examples/getting_started.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "08f9d077-64ed-4dc3-b772-498569e96d2b",
   "metadata": {
    "tags": []
   },
   "source": [
    "# `mle-hyperopt`: Lightweight Hyperparameter Optimization 🚂\n",
    "### Author: [@RobertTLange](https://twitter.com/RobertTLange) [Last Update: March 2023][![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/mle-infrastructure/mle-hyperopt/blob/main/examples/getting_started.ipynb)\n",
    "\n",
    "Validating a simulation across a large range of parameters or tuning the hyperparameters of a neural network is common practice for every computational scientist. There are a plethora of open source tools that implement individual algorithms, but many of them are either combersome to set up and log or follow diverse syntax, which makes it hard to easily wrap them. For my personal research setup I wanted a simple API that allows me to generate batches of parameter configurations for various types of experiments and that comes with a set of handy features. These included the following: \n",
    "\n",
    "- **API simplicity**: The `mle-hyperopt` uses a simple interface alá `strategy.ask()` and `strategy.tell()`. Furthermore, many of the search spaces can be exported and repurposed. E.g. if you want to autovectorize/`jax.vmap` over a grid of parameters.\n",
    "- **Strategy diversity**: While most hyperparameter tools implement the newest population-based search algorithms and Bayesian Optimization variants, often times they do not feature simple grid search. Even though [Bergstra and Bengio (2012)](https://www.jmlr.org/papers/volume13/bergstra12a/bergstra12a.pdf) showed that random search can be more efficient, often an intuitive understanding of a grey-box model can be enhanced by semi-exhaustive evaluation.\n",
    "- **Interactive search space refinement**: After a certain set of search iterations, it can make sense to refine your search space boundaries based on the top performing configurations. This way the search can focus on a smaller range of promising configurations. This was for example done in [Schmidt et al. (2021)](https://arxiv.org/abs/2007.01547) and significantly improve computation efficient resource allocation. \n",
    "- **Exporting of configuration files**: Often times I want to submit a batch of training runs to a compute cluster (Slurm, etc.) and need to execute a downstream training routine that looks somewhat as follows: `python train.py --config_fname config.yaml`. In that case it is useful to have one script that executes calls to the cluster scheduler after having generated the configurations.\n",
    "- **Storage and reloading of previous search logs**: I was astonished by how few libraries provide the simple utility of exporting and importing a previous search experiment for later continuation. E.g., in order to accomplish something similar in the FAIR's awesome [nevergrad](https://facebookresearch.github.io/nevergrad/) library, one has to write manual functionality that dumps the search log, reloads and supplies the previously stored results to a new search strategy instance.\n",
    "\n",
    "And this was how the `mle-hyperopt` package was born. [***Note***: I by no means claim that this is something novel. Most likely you may your own substitute tool. But maybe you find a couple of the package's features useful. So hang in there 🤗] As of writing the package includes a set of diverse (e.g. single vs. multi-objective, model-free vs. model-based) search algorithms:\n",
    "\n",
    "| | Search Type           | Description | `search_config` |\n",
    "|----|----------------------- | ----------- | --------------- |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/grid.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `GridSearch`          |  Search over list of discrete values  | - |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/random.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `RandomSearch`        |  Random search over variable ranges         | `refine_after`, `refine_top_k` |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/coordinate.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `CoordinateSearch`    |  Coordinate-wise optimization with fixed defaults | `order`, `defaults`\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/smbo.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `SMBOSearch`          |  Sequential model-based optimization [(Hutter et al., 2011)](https://ml.informatik.uni-freiburg.de/wp-content/uploads/papers/11-LION5-SMAC.pdf)   | `base_estimator`, `acq_function`, `n_initial_points`\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/nevergrad.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `NevergradSearch`     |  Multi-objective [nevergrad](https://facebookresearch.github.io/nevergrad/) wrapper | `optimizer`, `budget_size`, `num_workers`\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/halving.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `HalvingSearch`     | Successive Halving [(Karmin et al., 2013)](https://proceedings.mlr.press/v28/karnin13.html) | `min_budget`, `num_arms`, `halving_coeff`\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/hyperband.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `HyperbandSearch`     | Hyperband [(Li et al., 2018)](https://arxiv.org/pdf/1603.06560.pdf) | `max_resource`, `eta`\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/pbt.png?raw=true\" alt=\"drawing\" width=\"65\"/>|  `PBTSearch`     | Population-Based Training [(Jaderberg et al., 2017)](https://arxiv.org/pdf/1711.09846.pdf) | `explore`, `exploit`\n",
    "\n",
    "Each strategy implements a separate search space which one can sample from. The API follows the standard `ask`, `eval`, `tell` paradigm. The general package structure is summarized as follows:\n",
    "\n",
    "<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/mle_hyperopt_structure.png?raw=true\" alt=\"drawing\" width=\"900\"/>\n",
    "\n",
    "In the remainder of this 'blog notebook' we will walk through the different features and use-cases. Let's start by implementing a small 'synthetic' helper function that evaluates the performance of a combination of 3 standard hyperparameters -- learning rate, batchsize and architecture:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b1ff5d35-b29e-4386-b423-9c83ad1ae169",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "try:\n",
    "    import mle_hyperopt\n",
    "except:\n",
    "    !pip install -q mle-hyperopt[examples]\n",
    "    import mle_hyperopt\n",
    "\n",
    "def fake_train(lrate, batch_size, arch):\n",
    "    \"\"\"Optimum: lrate=0.2, batch_size=4, arch='conv'.\"\"\"\n",
    "    f1 = ((lrate - 0.2) ** 2 + ((batch_size - 4)/4) ** 2\n",
    "          + (0 if arch == \"conv\" else 0.2))\n",
    "    return f1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b84d6078-40a7-4715-89db-927b0e380025",
   "metadata": {},
   "source": [
    "## Basic API Usage: Grid Search <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/grid.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "Imagine you simply want to loop over a discretized range of hyperparameters for our training surrogate. In order to do so we first need to define the search space. The `mle-hyperopt` package supports real-, integer- and categorically-valued parameters, whose ranges you specify via dictionaries. For real variables and integers you have to specifiy the beginning and end of the range (`begin`/`end`) as well as a prior (e.g. uniform or log-uniform) or the number of bins to discretize (`prior`/`bins`). For categorical variables simply supply a list of values:\n",
    "\n",
    "| | Variable            | Type | Space Specification |\n",
    "| --- |----------------------- | ----------- | --------------- |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/real.png?raw=true\" alt=\"drawing\" width=\"50\"/> |  **`real`**          |  Real-valued  | `Dict`: `begin`, `end`, `prior`/`bins` (grid) |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/integer.png?raw=true\" alt=\"drawing\" width=\"50\"/>  |  **`integer`**        |  Integer-valued         | `Dict`: `begin`, `end`, `prior`/`bins` (grid) |\n",
    "|<img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/categorical.png?raw=true\" alt=\"drawing\" width=\"50\"/> |  **`categorical`**  |  Categorical        | `List`: Values to search over\n",
    "\n",
    "Let's now instantiate of search strategy for a range of learning rates and batch sizes as well as two network types we want to evaluate:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "19c7488b-34a3-47fc-837a-f088cfaa48ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">         <span style=\"font-style: italic\">               MLE-Hyperopt Grid Hyperspace 🚀               </span>          \n",
       "                                                                                \n",
       "          <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                </span>           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['mlp', 'cnn']                </span>           \n",
       "           lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.1, End: 0.5, Bins: 5 </span>           \n",
       "           batch_size    integer      <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 1, End: 5, Bins: 1     </span>           \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "         \u001b[3m               MLE-Hyperopt Grid Hyperspace 🚀               \u001b[0m          \n",
       "                                                                                \n",
       "          \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔               \u001b[0m\u001b[1;31m \u001b[0m           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           arch          categorical  \u001b[31m \u001b[0m\u001b[31m['mlp', 'cnn']               \u001b[0m\u001b[31m \u001b[0m           \n",
       "           lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.1, End: 0.5, Bins: 5\u001b[0m\u001b[31m \u001b[0m           \n",
       "           batch_size    integer      \u001b[31m \u001b[0m\u001b[31mBegin: 1, End: 5, Bins: 1    \u001b[0m\u001b[31m \u001b[0m           \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:44] </span>Start running 3D grid with <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">10</span> total configurations.      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#163\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">163</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:44]\u001b[0m\u001b[2;36m \u001b[0mStart running 3D grid with \u001b[1;36m10\u001b[0m total configurations.      \u001b]8;id=748486;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=280396;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#163\u001b\\\u001b[2m163\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mle_hyperopt import GridSearch\n",
    "# Instantiate grid search class\n",
    "strategy = GridSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                      \"end\": 0.5,\n",
    "                                      \"bins\": 5}},\n",
    "                      integer={\"batch_size\": {\"begin\": 1,\n",
    "                                              \"end\": 5,\n",
    "                                              \"bins\": 1}},\n",
    "                      categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                      verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e7bfa735-12ed-4b88-b0ce-76f5ccb5ab1a",
   "metadata": {},
   "source": [
    "If you set the verbosity option, the logged data will be printed to your console using the beautiful [rich](https://github.com/willmcgugan/rich) package. We are now ready to ask our search strategy for a set of proposal candidates:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "49111004-e317-4b1e-ad63-8e32ccf8ecc1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'arch': 'mlp', 'batch_size': 1, 'lrate': 0.1},\n",
       " {'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2},\n",
       " {'arch': 'mlp', 'batch_size': 1, 'lrate': 0.30000000000000004},\n",
       " {'arch': 'mlp', 'batch_size': 1, 'lrate': 0.4},\n",
       " {'arch': 'mlp', 'batch_size': 1, 'lrate': 0.5}]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Ask the strategy for five configs to evaluate\n",
    "configs = strategy.ask(batch_size=5)\n",
    "configs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3985f7cd-8f04-4bb0-bb99-9030d4f339de",
   "metadata": {},
   "source": [
    "Next, we can quickly evaluate our fake surrogate objective for the two proposal configurations and afterwards update the strategy with the collected data (e.g. a validation score). Note that the API assumes that we are minimizing an objective. If you want to instead maximize simply provide the option `maximize_objective = True` when instantiating the search strategy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "057f6ced-c142-4a42-93d9-cc58246d14df",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 5   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 12:55:48       </span>┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 5  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 12:55:48      \u001b[0m\u001b[1m \u001b[0m┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Simple ask - eval - tell API\n",
    "values = [fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c69df333-f2d6-4479-b588-23966b459a9a",
   "metadata": {},
   "source": [
    "The log with the individual search evaluation results can be stored and reloaded using `strategy.save()` and `strategy.load()`. You can store the strategy data as a json, yaml of pkl file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6be549ca-4062-4a33-b40c-b0a7217bc859",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span>Stored <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> search iterations ➞ search_log.json.                      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#354\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">354</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0mStored \u001b[1;36m5\u001b[0m search iterations ➞ search_log.json.                      \u001b]8;id=133144;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=967994;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#354\u001b\\\u001b[2m354\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span><span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'arch'</span>: <span style=\"color: #008000; text-decoration-color: #008000\">'mlp'</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'batch_size'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'lrate'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.1</span><span style=\"font-weight: bold\">}</span> was previously      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">261</span></a>\n",
       "<span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>evaluated.                                                         <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0m\u001b[1m{\u001b[0m\u001b[32m'arch'\u001b[0m: \u001b[32m'mlp'\u001b[0m, \u001b[32m'batch_size'\u001b[0m: \u001b[1;36m1\u001b[0m, \u001b[32m'lrate'\u001b[0m: \u001b[1;36m0.1\u001b[0m\u001b[1m}\u001b[0m was previously      \u001b]8;id=994129;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=499359;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\u001b\\\u001b[2m261\u001b[0m\u001b]8;;\u001b\\\n",
       "\u001b[2;36m           \u001b[0mevaluated.                                                         \u001b[2m               \u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span><span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'arch'</span>: <span style=\"color: #008000; text-decoration-color: #008000\">'mlp'</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'batch_size'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'lrate'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.2</span><span style=\"font-weight: bold\">}</span> was previously      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">261</span></a>\n",
       "<span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>evaluated.                                                         <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0m\u001b[1m{\u001b[0m\u001b[32m'arch'\u001b[0m: \u001b[32m'mlp'\u001b[0m, \u001b[32m'batch_size'\u001b[0m: \u001b[1;36m1\u001b[0m, \u001b[32m'lrate'\u001b[0m: \u001b[1;36m0.2\u001b[0m\u001b[1m}\u001b[0m was previously      \u001b]8;id=186051;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=701540;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\u001b\\\u001b[2m261\u001b[0m\u001b]8;;\u001b\\\n",
       "\u001b[2;36m           \u001b[0mevaluated.                                                         \u001b[2m               \u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span><span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'arch'</span>: <span style=\"color: #008000; text-decoration-color: #008000\">'mlp'</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'batch_size'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'lrate'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.30000000000000004</span><span style=\"font-weight: bold\">}</span> was <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">261</span></a>\n",
       "<span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>previously evaluated.                                              <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0m\u001b[1m{\u001b[0m\u001b[32m'arch'\u001b[0m: \u001b[32m'mlp'\u001b[0m, \u001b[32m'batch_size'\u001b[0m: \u001b[1;36m1\u001b[0m, \u001b[32m'lrate'\u001b[0m: \u001b[1;36m0.30000000000000004\u001b[0m\u001b[1m}\u001b[0m was \u001b]8;id=442154;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=395021;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\u001b\\\u001b[2m261\u001b[0m\u001b]8;;\u001b\\\n",
       "\u001b[2;36m           \u001b[0mpreviously evaluated.                                              \u001b[2m               \u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span><span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'arch'</span>: <span style=\"color: #008000; text-decoration-color: #008000\">'mlp'</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'batch_size'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'lrate'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.4</span><span style=\"font-weight: bold\">}</span> was previously      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">261</span></a>\n",
       "<span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>evaluated.                                                         <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0m\u001b[1m{\u001b[0m\u001b[32m'arch'\u001b[0m: \u001b[32m'mlp'\u001b[0m, \u001b[32m'batch_size'\u001b[0m: \u001b[1;36m1\u001b[0m, \u001b[32m'lrate'\u001b[0m: \u001b[1;36m0.4\u001b[0m\u001b[1m}\u001b[0m was previously      \u001b]8;id=616219;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=10463;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\u001b\\\u001b[2m261\u001b[0m\u001b]8;;\u001b\\\n",
       "\u001b[2;36m           \u001b[0mevaluated.                                                         \u001b[2m               \u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span><span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'arch'</span>: <span style=\"color: #008000; text-decoration-color: #008000\">'mlp'</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'batch_size'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'lrate'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.5</span><span style=\"font-weight: bold\">}</span> was previously      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">261</span></a>\n",
       "<span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>evaluated.                                                         <span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0m\u001b[1m{\u001b[0m\u001b[32m'arch'\u001b[0m: \u001b[32m'mlp'\u001b[0m, \u001b[32m'batch_size'\u001b[0m: \u001b[1;36m1\u001b[0m, \u001b[32m'lrate'\u001b[0m: \u001b[1;36m0.5\u001b[0m\u001b[1m}\u001b[0m was previously      \u001b]8;id=288773;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=372801;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#261\u001b\\\u001b[2m261\u001b[0m\u001b]8;;\u001b\\\n",
       "\u001b[2;36m           \u001b[0mevaluated.                                                         \u001b[2m               \u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:51] </span>Reloaded <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0</span> previous search iterations.                             <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#416\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">416</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:51]\u001b[0m\u001b[2;36m \u001b[0mReloaded \u001b[1;36m0\u001b[0m previous search iterations.                             \u001b]8;id=465562;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=837607;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#416\u001b\\\u001b[2m416\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Storing of results to .pkl\n",
    "strategy.save(\"search_log.json\")\n",
    "\n",
    "# Reloading of results from .pkl\n",
    "strategy.load(\"search_log.json\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f76caceb-fec1-4883-b288-5b1eca5ba4a0",
   "metadata": {},
   "source": [
    "Note that no configuration is added to the log since they were already previously archived. This is different if we instatiate a new strategy with the `reload_path` option. In this case the new search strategy will load the previously stored log from the file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d0309087-669c-41c6-b450-0275b6680459",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:55:53] </span>Reloaded <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> previous search iterations.                             <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">strategy.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#416\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">416</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:55:53]\u001b[0m\u001b[2;36m \u001b[0mReloaded \u001b[1;36m5\u001b[0m previous search iterations.                             \u001b]8;id=931973;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py\u001b\\\u001b[2mstrategy.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=833106;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategy.py#416\u001b\\\u001b[2m416\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mle_hyperopt import GridSearch\n",
    "strategy = GridSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                      \"end\": 0.5,\n",
    "                                      \"bins\": 5}},\n",
    "                      integer={\"batch_size\": {\"begin\": 1,\n",
    "                                              \"end\": 5,\n",
    "                                              \"bins\": 1}},\n",
    "                      categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                      reload_path=\"search_log.json\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aae59ff3-7b18-47eb-bf34-18565db7e13b",
   "metadata": {},
   "source": [
    "There are different ways how you can access and analyse the search results:\n",
    "\n",
    "- `strategy.log` will return the raw internal list of evaluation results.\n",
    "- `strategy.df` will return the flattened log as a pandas dataframe.\n",
    "- `strategy.get_best(top_k)` will retrieve the `top_k` best performing configurations.\n",
    "\n",
    "But my two favorite tools are a plot of the best performance over search iterations and the overall ranking:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b7490e4d-bf23-491e-a603-229b6a52add8",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<frozen importlib._bootstrap>:228: RuntimeWarning: scipy._lib.messagestream.MessageStream size changed, may indicate binary incompatibility. Expected 56 from C header, got 64 from PyObject\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(<Figure size 432x288 with 1 Axes>,\n",
       " <Axes: title={'center': 'Best Objective Value'}, xlabel='# Config Evaluations', ylabel='Objective'>)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 280,
       "width": 425
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot timeseries of best performing score over search iterations\n",
    "strategy.plot_best()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e7495374-6249-4db8-930c-86ed6d664b96",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">   ┏━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓    \n",
       "   ┃<span style=\"font-weight: bold\"> 🥇 Rank </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖                             </span>┃    \n",
       "   ┡━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩    \n",
       "   │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 1       </span>│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │    \n",
       "   │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 2       </span>│ 0  │ 0.772   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.1 │    \n",
       "   │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 3       </span>│ 2  │ 0.772   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.3 │    \n",
       "   │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 4       </span>│ 3  │ 0.802   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.4 │    \n",
       "   └─────────┴────┴─────────┴──────────────────────────────────────────────┘    \n",
       "</pre>\n"
      ],
      "text/plain": [
       "   ┏━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓    \n",
       "   ┃\u001b[1m \u001b[0m\u001b[1m🥇 Rank\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖                            \u001b[0m\u001b[1m \u001b[0m┃    \n",
       "   ┡━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩    \n",
       "   │\u001b[2m \u001b[0m\u001b[2m1      \u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.762   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.2 │    \n",
       "   │\u001b[2m \u001b[0m\u001b[2m2      \u001b[0m\u001b[2m \u001b[0m│ 0  │ 0.772   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.1 │    \n",
       "   │\u001b[2m \u001b[0m\u001b[2m3      \u001b[0m\u001b[2m \u001b[0m│ 2  │ 0.772   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.3 │    \n",
       "   │\u001b[2m \u001b[0m\u001b[2m4      \u001b[0m\u001b[2m \u001b[0m│ 3  │ 0.802   │ 'arch': 'mlp', 'batch_size': 1, 'lrate': 0.4 │    \n",
       "   └─────────┴────┴─────────┴──────────────────────────────────────────────┘    \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Print out ranking of best performers\n",
    "strategy.print_ranking(top_k=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "289a9690-f708-4c15-847c-73a9b2be6abc",
   "metadata": {},
   "source": [
    "In case you don't want to search over all configuration parameters, you can also add a dictionary of fixed parameters at instantiation (`fixed_params`). Additionally, you can also directly store the configurations as files by setting `store=True` (or also provide the filenames as a list using the `config_fnames` option). Below is an example for the case where we want to add a `momentum` parameter with value 0.9 to all configurations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d312aece-7a41-4ada-84e2-c0c79497af30",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([{'lrate': 0.1, 'arch': 'mlp', 'batch_size': 1, 'momentum': 0.9},\n",
       "  {'lrate': 0.2, 'arch': 'mlp', 'batch_size': 1, 'momentum': 0.9}],\n",
       " ['eval_0.yaml', 'eval_1.yaml'])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "strategy = GridSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                      \"end\": 0.5,\n",
    "                                      \"bins\": 5}},\n",
    "                      integer={\"batch_size\": {\"begin\": 1,\n",
    "                                              \"end\": 5,\n",
    "                                              \"bins\": 1}},\n",
    "                      categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                      fixed_params={\"momentum\": 0.9})\n",
    "\n",
    "strategy.ask(2, store=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e861c990-2497-44a7-a956-a44be45eadc6",
   "metadata": {},
   "source": [
    "## Random Search & Search Space Refinement <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/random.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "Next let's see how we can define a random search strategy with priors over variables. As before we will define our search space using a set of dictionaries, but now the real and integer parameters will have a uniform or log-uniform prior placed over their range:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0d9c2fa2-2941-49ca-9f34-49b9a0e53b0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">      <span style=\"font-style: italic\">                 MLE-Hyperopt Random Hyperspace 🚀                  </span>      \n",
       "                                                                                \n",
       "       <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                       </span>       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['mlp', 'cnn']                       </span>       \n",
       "        lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.1, End: 0.5, Prior: uniform </span>       \n",
       "        batch_size    integer      <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 1, End: 5, Prior: log-uniform </span>       \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "      \u001b[3m                 MLE-Hyperopt Random Hyperspace 🚀                  \u001b[0m      \n",
       "                                                                                \n",
       "       \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                      \u001b[0m\u001b[1;31m \u001b[0m       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  \u001b[31m \u001b[0m\u001b[31m['mlp', 'cnn']                      \u001b[0m\u001b[31m \u001b[0m       \n",
       "        lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.1, End: 0.5, Prior: uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "        batch_size    integer      \u001b[31m \u001b[0m\u001b[31mBegin: 1, End: 5, Prior: log-uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 5   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 12:55:58        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 4  │ 0.268   │ 'arch': 'cnn', 'lrate': 0.123, 'batch_size':  │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 3                                             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 4  │ 0.268   │ 'arch': 'cnn', 'lrate': 0.123, 'batch_size':  │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 3                                             │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 5  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 12:55:58       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 4  │ 0.268   │ 'arch': 'cnn', 'lrate': 0.123, 'batch_size':  │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 3                                             │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 4  │ 0.268   │ 'arch': 'cnn', 'lrate': 0.123, 'batch_size':  │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 3                                             │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-style: italic\">         MLE-Hyperopt Random - 5 Evals - Top 2 - Refined Hyperspace 🚀          </span>\n",
       "                                                                                \n",
       " <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                                   </span> \n",
       " ────────────────────────────────────────────────────────────────────────────── \n",
       "  arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['cnn', 'mlp']                                   </span> \n",
       "  lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.123, End: 0.34044600469728353, Prior:   </span> \n",
       "                             <span style=\"color: #800000; text-decoration-color: #800000\"> uniform                                          </span> \n",
       "  batch_size    integer      <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 3, End: 3, Prior: log-uniform             </span> \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[3m         MLE-Hyperopt Random - 5 Evals - Top 2 - Refined Hyperspace 🚀          \u001b[0m\n",
       "                                                                                \n",
       " \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                                  \u001b[0m\u001b[1;31m \u001b[0m \n",
       " ────────────────────────────────────────────────────────────────────────────── \n",
       "  arch          categorical  \u001b[31m \u001b[0m\u001b[31m['cnn', 'mlp']                                  \u001b[0m\u001b[31m \u001b[0m \n",
       "  lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.123, End: 0.34044600469728353, Prior:  \u001b[0m\u001b[31m \u001b[0m \n",
       "                             \u001b[31m \u001b[0m\u001b[31muniform                                         \u001b[0m\u001b[31m \u001b[0m \n",
       "  batch_size    integer      \u001b[31m \u001b[0m\u001b[31mBegin: 3, End: 3, Prior: log-uniform            \u001b[0m\u001b[31m \u001b[0m \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mle_hyperopt import RandomSearch\n",
    "\n",
    "strategy = RandomSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                        \"end\": 0.5,\n",
    "                                        \"prior\": \"uniform\"}},\n",
    "                        integer={\"batch_size\": {\"begin\": 1,\n",
    "                                                \"end\": 5,\n",
    "                                                \"prior\": \"log-uniform\"}},\n",
    "                        categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                        search_config={\"refine_after\": 5,\n",
    "                                       \"refine_top_k\": 2},\n",
    "                        seed_id=42,\n",
    "                        verbose=True)\n",
    "\n",
    "configs = strategy.ask(5)\n",
    "values = [fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d936ba9a-4bd7-4a8d-bb36-7cb9cadb46cb",
   "metadata": {},
   "source": [
    "In the random search instantiation above we have additionally set the random seed for reproducibility and supplied a `search_config` dictionary with two keys: `refine_after` and `refine_top_k`. These two define when and how the search space will be contracted. More specifically, the ranges and categorical values will be altered to only include those of the top-k best performing configurations stored in the strategy log. This allows the strategy to spend more time sampling from regions of the search space, which have proven well-performing."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cff0d18-988b-4b57-a74e-0191e1834b4b",
   "metadata": {},
   "source": [
    "## SMBO <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/smbo.png?raw=true\" alt=\"drawing\" width=\"50\"/> & Coordinate-Wise Search <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/coordinate.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "The `mle-hyperopt` package also wraps around the [`scikit-optimize`](https://scikit-optimize.github.io/stable/index.html) API and provides a simple interface for SMBO ([Hutter et al., 2011](https://ml.informatik.uni-freiburg.de/wp-content/uploads/papers/11-LION5-SMAC.pdf)). You can customize the optimizer object with the `search_config` options. This includes different surrogate models (\"GP\", \"RF\", \"ET\", \"GBRT\") and acquisition functions (\"LCB\", \"EI\", \"PI\") as well as random initial seeding configurations. For more information check out the scikit optimizer [documentation](https://scikit-optimize.github.io/stable/modules/generated/skopt.Optimizer.html#skopt.Optimizer)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b72b435e-5141-4389-a0ae-4de9f1d60f0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mle_hyperopt import SMBOSearch\n",
    "\n",
    "strategy = SMBOSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                      \"end\": 0.5,\n",
    "                                      \"prior\": \"uniform\"}},\n",
    "                      integer={\"batch_size\": {\"begin\": 1,\n",
    "                                              \"end\": 5,\n",
    "                                              \"prior\": \"uniform\"}},\n",
    "                      search_config={\"base_estimator\": \"GP\",\n",
    "                                     \"acq_function\": \"gp_hedge\",\n",
    "                                     \"n_initial_points\": 5},\n",
    "                      fixed_params={\"arch\": \"cnn\"})\n",
    "\n",
    "configs = strategy.ask(5)\n",
    "values = [fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7a07dae-cc95-4ad0-af14-f30a672a12c0",
   "metadata": {},
   "source": [
    "Furthermore, I also implemented a more heuristic search, which I often would implement manually: We start by scanning one parameter and hold the other considered parameters fixed to a pre-specified `defaults` value. Afterwards, the previously scanned parameter is in turn fixed to its best value and we go over to the next parameter. We repeat this coordinate-wise search until all specified parameters are completed. The order in which we search over the variables is specified in the `search_config`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b078c46b-71b7-4891-93c6-56651ede3961",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:01] </span>New active variable/coordinate `lrate`.                          <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">coordinate.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">163</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:01]\u001b[0m\u001b[2;36m \u001b[0mNew active variable/coordinate `lrate`.                          \u001b]8;id=417868;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\u001b\\\u001b[2mcoordinate.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=955618;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\u001b\\\u001b[2m163\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">         <span style=\"font-style: italic\">            MLE-Hyperopt Coordinate Hyperspace 🚀            </span>          \n",
       "                                                                                \n",
       "          <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                </span>           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           batch_size    categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> [3]                           </span>           \n",
       "           arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['mlp']                       </span>           \n",
       "           lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.1, End: 0.5, Bins: 5 </span>           \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "         \u001b[3m            MLE-Hyperopt Coordinate Hyperspace 🚀            \u001b[0m          \n",
       "                                                                                \n",
       "          \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔               \u001b[0m\u001b[1;31m \u001b[0m           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           batch_size    categorical  \u001b[31m \u001b[0m\u001b[31m[3]                          \u001b[0m\u001b[31m \u001b[0m           \n",
       "           arch          categorical  \u001b[31m \u001b[0m\u001b[31m['mlp']                      \u001b[0m\u001b[31m \u001b[0m           \n",
       "           lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.1, End: 0.5, Bins: 5\u001b[0m\u001b[31m \u001b[0m           \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[{'arch': 'mlp', 'batch_size': 3, 'lrate': 0.1},\n",
       " {'arch': 'mlp', 'batch_size': 3, 'lrate': 0.2},\n",
       " {'arch': 'mlp', 'batch_size': 3, 'lrate': 0.30000000000000004},\n",
       " {'arch': 'mlp', 'batch_size': 3, 'lrate': 0.4},\n",
       " {'arch': 'mlp', 'batch_size': 3, 'lrate': 0.5}]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from mle_hyperopt import CoordinateSearch\n",
    "\n",
    "strategy = CoordinateSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                            \"end\": 0.5,\n",
    "                                            \"bins\": 5}},\n",
    "                            integer={\"batch_size\": {\"begin\": 1,\n",
    "                                                    \"end\": 5,\n",
    "                                                    \"bins\": 5}},\n",
    "                            categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                            search_config={\"order\": [\"lrate\", \"batch_size\", \"arch\"],\n",
    "                                           \"defaults\": {\"lrate\": 0.1,\n",
    "                                                        \"batch_size\": 3,\n",
    "                                                        \"arch\": \"mlp\"}},\n",
    "                            verbose=True)\n",
    "\n",
    "configs = strategy.ask(5)\n",
    "configs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "5191113d-debe-401f-b1a7-45b9a6626364",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 5   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 12:56:02       </span>┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 1  │ 0.262   │ 'arch': 'mlp', 'batch_size': 3, 'lrate': 0.2 │ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 1  │ 0.262   │ 'arch': 'mlp', 'batch_size': 3, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 5  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 12:56:02      \u001b[0m\u001b[1m \u001b[0m┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.262   │ 'arch': 'mlp', 'batch_size': 3, 'lrate': 0.2 │ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.262   │ 'arch': 'mlp', 'batch_size': 3, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:02] </span>Fixed `lrate` hyperparameter to <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.2</span>.                             <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">coordinate.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#158\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">158</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:02]\u001b[0m\u001b[2;36m \u001b[0mFixed `lrate` hyperparameter to \u001b[1;36m0.2\u001b[0m.                             \u001b]8;id=537618;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\u001b\\\u001b[2mcoordinate.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=481075;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#158\u001b\\\u001b[2m158\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:02] </span>New active variable/coordinate `batch_size`.                     <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">coordinate.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">163</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:02]\u001b[0m\u001b[2;36m \u001b[0mNew active variable/coordinate `batch_size`.                     \u001b]8;id=88929;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\u001b\\\u001b[2mcoordinate.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=406873;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\u001b\\\u001b[2m163\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "values = [fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "bb1e37cf-d2c2-42e9-8b86-71c335596849",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 9   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 12:56:02       </span>┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 9  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 12:56:02      \u001b[0m\u001b[1m \u001b[0m┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:02] </span>Fixed `batch_size` hyperparameter to <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>.                          <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">coordinate.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#158\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">158</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:02]\u001b[0m\u001b[2;36m \u001b[0mFixed `batch_size` hyperparameter to \u001b[1;36m4\u001b[0m.                          \u001b]8;id=241043;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\u001b\\\u001b[2mcoordinate.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=931582;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#158\u001b\\\u001b[2m158\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:02] </span>New active variable/coordinate `arch`.                           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">coordinate.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">163</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:02]\u001b[0m\u001b[2;36m \u001b[0mNew active variable/coordinate `arch`.                           \u001b]8;id=701366;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py\u001b\\\u001b[2mcoordinate.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=193043;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/strategies/coordinate.py#163\u001b\\\u001b[2m163\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "configs = strategy.ask(4)\n",
    "values = [fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1d95d587-acf3-46d0-bf1e-751dcc279c23",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 10  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 12:56:03       </span>┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 9  │ 0.2     │ 'arch': 'cnn', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ \n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 10 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 12:56:03      \u001b[0m\u001b[1m \u001b[0m┃ \n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 7  │ 0.2     │ 'arch': 'mlp', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 9  │ 0.2     │ 'arch': 'cnn', 'batch_size': 4, 'lrate': 0.2 │ \n",
       "└───────────────┴────┴─────────┴──────────────────────────────────────────────┘ \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "configs = strategy.ask()\n",
    "values = fake_train(**configs)\n",
    "strategy.tell(configs, values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "c0c850bc-9832-4598-a6b4-33ee133fb1c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<Figure size 432x288 with 1 Axes>,\n",
       " <Axes: title={'center': 'Best Objective Value'}, xlabel='# Config Evaluations', ylabel='Objective'>)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 402
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "strategy.plot_best()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04f28a40-32f2-4662-8e39-801019a5a03e",
   "metadata": {},
   "source": [
    "## Multi-Objective Hyperparameter Optimization with `nevergrad` <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/nevergrad.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "So far we have discussed a set of single objective hyperparameter search strategies. But how about cases where we are interested in more than one objective? E.g. test accuracy and inference time. For these cases and for even more diverse search strategies we wrap around the `nevergrad` library by Facebook Research. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "e85b4e98-0c67-4478-bae7-207203378831",
   "metadata": {},
   "outputs": [],
   "source": [
    "def multi_fake_train(lrate, batch_size, arch):\n",
    "    # optimal for learning_rate=0.2, batch_size=4, architecture=\"conv\"\n",
    "    f1 = ((lrate - 0.2) ** 2 + (batch_size - 4) ** 2\n",
    "          + (0 if arch == \"conv\" else 10))\n",
    "    # optimal for learning_rate=0.3, batch_size=2, architecture=\"mlp\"\n",
    "    f2 = ((lrate - 0.3) ** 2 + (batch_size - 2) ** 2\n",
    "          + (0 if arch == \"mlp\" else 5))\n",
    "    return f1, f2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "28631203-c216-4f71-aa58-038c83c7db3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'eval_id': 0,\n",
       "  'params': {'lrate': 0.3, 'batch_size': 4},\n",
       "  'objective': (10.01, 9.0)},\n",
       " {'eval_id': 1,\n",
       "  'params': {'lrate': 0.2890150583350444, 'batch_size': 4},\n",
       "  'objective': (10.007923680610391, 9.000120668943381)},\n",
       " {'eval_id': 2,\n",
       "  'params': {'lrate': 0.2644171178765383, 'batch_size': 2},\n",
       "  'objective': (14.004149565075519, 5.001266141500212)},\n",
       " {'eval_id': 3,\n",
       "  'params': {'lrate': 0.3768394539854681, 'batch_size': 3},\n",
       "  'objective': (11.03127219248588, 6.005904301688785)},\n",
       " {'eval_id': 4,\n",
       "  'params': {'lrate': 0.3, 'batch_size': 3},\n",
       "  'objective': (11.01, 6.0)}]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from mle_hyperopt import NevergradSearch\n",
    "\n",
    "strategy = NevergradSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                      \"end\": 0.5,\n",
    "                                      \"prior\": \"uniform\"}},\n",
    "                           integer={\"batch_size\": {\"begin\": 1,\n",
    "                                                   \"end\": 5,\n",
    "                                                   \"prior\": \"uniform\"}},\n",
    "                           search_config={\"optimizer\": \"NGOpt\",\n",
    "                                          \"budget_size\": 100,\n",
    "                                          \"num_workers\": 5},\n",
    "                           fixed_params={\"arch\": \"cnn\"})\n",
    "\n",
    "configs = strategy.ask(5)\n",
    "values = [multi_fake_train(**c) for c in configs]\n",
    "strategy.tell(configs, values)\n",
    "strategy.log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8018e5ad-4e9c-4291-841e-2da05d1222a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">      ┏━━━━━━━━━┳━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓      \n",
       "      ┃<span style=\"font-weight: bold\"> 🥇 Rank </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉         </span>┃<span style=\"font-weight: bold\"> Configuration 🔖                </span>┃      \n",
       "      ┡━━━━━━━━━╇━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩      \n",
       "      │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 1       </span>│ -  │ [10.008, 9.0]   │ 'lrate': 0.289, 'batch_size': 4 │      \n",
       "      │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 2       </span>│ -  │ [10.01, 9.0]    │ 'lrate': 0.3, 'batch_size': 4   │      \n",
       "      │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 3       </span>│ -  │ [11.01, 6.0]    │ 'lrate': 0.3, 'batch_size': 3   │      \n",
       "      │<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> 4       </span>│ -  │ [14.004, 5.001] │ 'lrate': 0.264, 'batch_size': 2 │      \n",
       "      └─────────┴────┴─────────────────┴─────────────────────────────────┘      \n",
       "</pre>\n"
      ],
      "text/plain": [
       "      ┏━━━━━━━━━┳━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓      \n",
       "      ┃\u001b[1m \u001b[0m\u001b[1m🥇 Rank\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉        \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖               \u001b[0m\u001b[1m \u001b[0m┃      \n",
       "      ┡━━━━━━━━━╇━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩      \n",
       "      │\u001b[2m \u001b[0m\u001b[2m1      \u001b[0m\u001b[2m \u001b[0m│ -  │ [10.008, 9.0]   │ 'lrate': 0.289, 'batch_size': 4 │      \n",
       "      │\u001b[2m \u001b[0m\u001b[2m2      \u001b[0m\u001b[2m \u001b[0m│ -  │ [10.01, 9.0]    │ 'lrate': 0.3, 'batch_size': 4   │      \n",
       "      │\u001b[2m \u001b[0m\u001b[2m3      \u001b[0m\u001b[2m \u001b[0m│ -  │ [11.01, 6.0]    │ 'lrate': 0.3, 'batch_size': 3   │      \n",
       "      │\u001b[2m \u001b[0m\u001b[2m4      \u001b[0m\u001b[2m \u001b[0m│ -  │ [14.004, 5.001] │ 'lrate': 0.264, 'batch_size': 2 │      \n",
       "      └─────────┴────┴─────────────────┴─────────────────────────────────┘      \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "strategy.print_ranking()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d91a5267-ce18-4bc7-b426-814b0c36729f",
   "metadata": {},
   "source": [
    "# `hyperopt` decorator - minimal search wrapper\n",
    "\n",
    "For convenience I also added a simple `hyperopt` function decorator, which allows you to automatically execute the search procedure loop. The decorator assumes that the function to evaluate directly consumes a configuration dictionary:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "f19fefb1-63e2-43f9-a7b8-dcb56fc7ff4c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "400"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from mle_hyperopt import hyperopt\n",
    "\n",
    "@hyperopt(strategy_type=\"Grid\",\n",
    "          num_search_iters=400,\n",
    "          real={\"x\": {\"begin\": -0.5, \"end\": 0.5, \"bins\": 20},\n",
    "                \"y\": {\"begin\": -0.5, \"end\": 0.5, \"bins\": 20}})\n",
    "def circle_objective(config):\n",
    "    distance = abs((config[\"x\"] ** 2 + config[\"y\"] ** 2))\n",
    "    return distance\n",
    "\n",
    "strategy = circle_objective()\n",
    "len(strategy)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1474c8e-d9e8-499e-8ecf-6dd5b640f583",
   "metadata": {},
   "source": [
    "Additionally you can plot the results of any grid search as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "cbd4b727-21b6-4b80-9b59-2ccb37e58695",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-0.5        -0.44736842 -0.39473684 -0.34210526 -0.28947368 -0.23684211\n",
      " -0.18421053 -0.13157895 -0.07894737 -0.02631579  0.02631579  0.07894737\n",
      "  0.13157895  0.18421053  0.23684211  0.28947368  0.34210526  0.39473684\n",
      "  0.44736842  0.5       ] [-0.5        -0.44736842 -0.39473684 -0.34210526 -0.28947368 -0.23684211\n",
      " -0.18421053 -0.13157895 -0.07894737 -0.02631579  0.02631579  0.07894737\n",
      "  0.13157895  0.18421053  0.23684211  0.28947368  0.34210526  0.39473684\n",
      "  0.44736842  0.5       ]\n",
      "20\n",
      "20\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(<Figure size 720x576 with 2 Axes>,\n",
       " <Axes: title={'center': 'Quadratics for Life\\nHow beautiful can they be?'}, xlabel='x', ylabel='y'>)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 537,
       "width": 598
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "strategy.plot_grid(params_to_plot=[\"x\", \"y\"],\n",
    "                   target_to_plot=\"objective\",\n",
    "                   plot_title=\"Quadratics for Life\",\n",
    "                   plot_subtitle=\"How beautiful can they be?\",\n",
    "                   xy_labels= [\"x\", \"y\"],\n",
    "                   variable_name=\"Objective Name\",\n",
    "                   every_nth_tick=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e330db6f-c304-4d26-8e67-39033a7574f0",
   "metadata": {},
   "source": [
    "If there are more than two search variables please additionally provide a `fixed_params` dictionary specifying the variable name and value to fix it to in the plot."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "286df9ec-bdfa-4d3f-bb0b-549fcc21f43f",
   "metadata": {},
   "source": [
    "# Integration with `mle-logging` <img src=\"https://github.com/mle-infrastructure/mle-logging/blob/main/docs/logo_transparent.png?raw=true\" alt=\"drawing\" width=\"50\"/> & `mle-scheduler` <img src=\"https://github.com/mle-infrastructure/mle-scheduler/blob/main/docs/logo_transparent.png?raw=true\" alt=\"drawing\" width=\"50\"/>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "ea3edf7a-641f-4bda-abea-dcb86d38e55c",
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    from mle_logging import MLELogger\n",
    "except:\n",
    "    !pip install -q mle-logging\n",
    "    from mle_logging import MLELogger"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67666f03-b178-4ac2-8332-ee3e3296c3fb",
   "metadata": {},
   "source": [
    "The `mle-hyperopt` package also smoothly integrates with the [`mle-logging`](https://github.com/mle-infrastructure/mle-logging) package, which can be leveraged to log training statistics, checkpoints and other objects of interest. You can simply store the configurations and provide the configuration filename as an input to the `MLELogger`. The logger will then load the configuration and copy it to the logging directory:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2bc07c05-8785-4caa-a157-a8ed4c09de1d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">          <span style=\"font-style: italic\">              MLE-Hyperopt Grid Hyperspace 🚀              </span>           \n",
       "                                                                                \n",
       "           <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                     </span>            \n",
       "           ─────────────────────────────────────────────────────────            \n",
       "            lrate         real  <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 1e-08, End: 1e-06, Bins: 10 </span>            \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "          \u001b[3m              MLE-Hyperopt Grid Hyperspace 🚀              \u001b[0m           \n",
       "                                                                                \n",
       "           \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType\u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                    \u001b[0m\u001b[1;31m \u001b[0m            \n",
       "           ─────────────────────────────────────────────────────────            \n",
       "            lrate         real  \u001b[31m \u001b[0m\u001b[31mBegin: 1e-08, End: 1e-06, Bins: 10\u001b[0m\u001b[31m \u001b[0m            \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[12:56:20] </span>Start running 1D grid with <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">10</span> total configurations.      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#163\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">163</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[12:56:20]\u001b[0m\u001b[2;36m \u001b[0mStart running 1D grid with \u001b[1;36m10\u001b[0m total configurations.      \u001b]8;id=661622;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=253133;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#163\u001b\\\u001b[2m163\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'lrate': 1e-08}"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Store a set of configurations\n",
    "strategy = GridSearch(real={\"lrate\": {\"begin\": 1e-08, \"end\": 1e-06, \"bins\": 10}},\n",
    "                      verbose=True)\n",
    "configs, config_fnames = strategy.ask(3, store=True)\n",
    "\n",
    "# Instantiate logging to experiment_dir & pass configuration path\n",
    "log = MLELogger(time_to_track=['num_updates', 'num_epochs'],\n",
    "                what_to_track=['train_loss', 'test_loss'],\n",
    "                experiment_dir=\"experiment_dir/\",\n",
    "                config_fname=config_fnames[0],\n",
    "                use_tboard=False,\n",
    "                model_type='torch')\n",
    "\n",
    "log.update({'num_updates': 1, 'num_epochs': 1},\n",
    "           {'train_loss': 0.1, 'test_loss': 0.15},\n",
    "           save=True)\n",
    "\n",
    "# Check loaded and copied configuration\n",
    "log.config_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea2ad9d7-9a74-4a4d-949e-e76b83e2a14b",
   "metadata": {},
   "source": [
    "Furthermore, we can use the `mle-scheduler` package to directly execute the training script and aggregate `MLELogger` logs across multiple seeds. Below we train the weights of a small quadratic in PyTorch for multiple configurations in parallel:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "792fd65e-6b1b-4a13-ba59-4c1f20e3a738",
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    from mle_scheduler import MLEQueue\n",
    "except:\n",
    "    !pip install -q mle-scheduler\n",
    "    from mle_scheduler import MLEQueue\n",
    "    \n",
    "# Check if code is run in Colab: If so -- download py script from repo\n",
    "try:\n",
    "    import google.colab\n",
    "    IN_COLAB = True\n",
    "    !wget https://raw.githubusercontent.com/mle-infrastructure/mle-hyperopt/main/examples/iter_search/train_mlp.py\n",
    "    path_to_exec = \"train_mlp.py\"\n",
    "except:\n",
    "    IN_COLAB = False\n",
    "    path_to_exec = \"iter_search/train_mlp.py\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "649fa7cb-cf0e-4000-b78c-78bfac4c4fce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ba6053b3edf8413aa29f56fa9db03278",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"></pre>\n"
      ],
      "text/plain": []
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/rob/anaconda3/envs/mle-toolbox/lib/python3.9/site-packages/numpy/lib/function_base.py:4650: UserWarning: Warning: 'partition' will ignore the 'mask' of the MaskedArray.\n",
      "  arr.partition(\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 3   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:00:56        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 2  │ 17.157  │ 'lrate': 0.0, 'ckpt':                         │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 'logs/eval_2/models/final/final_seed_0.pt'    │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 2  │ 17.157  │ 'lrate': 0.0, 'ckpt':                         │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 'logs/eval_2/models/final/final_seed_0.pt'    │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 3  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:00:56       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 2  │ 17.157  │ 'lrate': 0.0, 'ckpt':                         │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 'logs/eval_2/models/final/final_seed_0.pt'    │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 2  │ 17.157  │ 'lrate': 0.0, 'ckpt':                         │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 'logs/eval_2/models/final/final_seed_0.pt'    │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "queue = MLEQueue(\n",
    "    resource_to_run=\"local\",\n",
    "    job_filename=path_to_exec,\n",
    "    config_filenames=config_fnames,\n",
    "    experiment_dir=\"logs\",\n",
    "    automerge_configs=True,\n",
    ")\n",
    "queue.run()\n",
    "\n",
    "scores = [queue.log[r].stats.loss.mean[-1] for r in queue.mle_run_ids]\n",
    "# You can also `tell` the strategy where you store the model ckpts\n",
    "ckpts = [queue.log[r].meta.model_ckpt for r in queue.mle_run_ids]\n",
    "strategy.tell(configs, scores, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55696d29-38a6-487e-987e-efa07350c7d3",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Advanced Search Strategies with Checkpoint Bookkeeping\n",
    "\n",
    "Until now all of the introduced search strategies rely on the full evaluation of a training run. Afterwards, we provide the strategy with the batch of perfomances for the different configurations and move on to the next batch -- that starts to train from scratch. Alternatively, there are different techniques, which make use of the iterative nature of most training procedures (e.g. anything gradient descent based). These training runs can easily be interrupted and either terminated or continued based on their performance up to now. Thereby, it is possible to allocate compute resources more efficiently by focussing on more promising candidates, while early stopping of bad ones. Importantly, this requires us to keep track of the checkpoints/network states of our different runs. Next we will have a look at the three most popular algorithms of that nature which are implemented in `mle-hyperopt`.\n",
    "\n",
    "## Successive Halving <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/halving.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "Successive Halving (SH), first proposed by [(Karmin et al., 2013)](https://proceedings.mlr.press/v28/karnin13.html), is a simple strategy which originated in the context of finding the best arm in a multi-arm bandit setting. Given a set of potential arms, `num_arms`, one starts by evaluating all of them for a limited number of pulls. Afterwards, the worst arms are discarded and one continues evaluating the best ones for more pulls. This process continues and we discard more and more arms, while spending more and more pulls on the most promising arms.\n",
    "\n",
    "In the context of tuning neural networks, the number of considered arms corresponds to the different configurations we are interested in evaluating. These can for example be randomly sampled. A \"pull\" from an arm then is analogous to training the network-configuration pair for a couple of steps. The aggressiveness with which we discard configurations is a hyperparameter (`halving_coeff`) and trades off **how many** configurations we can evaluate for **how long**. Finally, the strategy takes as an input the minimal number of pulls from all arms: `min_budget`. Let us again define a surrogate objective, which is \"iterative\" in its nature (taking `epoch` as an input):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "f2916ccf-21b3-4b3c-8b19-3f675854330e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "def get_iteration_score(\n",
    "    epoch: int, seed_id: int, lrate: float, batch_size: int, arch: str, **kwargs\n",
    ") -> (float, float):\n",
    "    \"\"\"Surrogate Objective w. optimum: lrate=0.2, batch_size=4, arch='conv'.\"\"\"\n",
    "    f1 = (lrate - 0.2) ** 2 + (batch_size - 4) ** 2 + (0 if arch == \"conv\" else 10)\n",
    "    train_loss = f1 + seed_id * 0.5\n",
    "    test_loss = f1 + seed_id * 0.5 + np.random.uniform(0, 0.3)\n",
    "    return train_loss / epoch, test_loss / epoch"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a097bde0-26dd-4422-9c51-398765b43538",
   "metadata": {},
   "source": [
    "Next, we instantiate the search class with a given set of hyperparameters. During the first `ask` call 20 configurations will be randomly sampled from the search space. Setting `halving_coeff` to 2 implies that at each `tell` call 50% of the still considered configurations will be discarded and that the amount of step iterations for the remaining ones is doubled:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "e74ae8e6-3b3c-4b02-b4fb-045e586be2aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">      <span style=\"font-style: italic\">                 MLE-Hyperopt Halving Hyperspace 🚀                 </span>      \n",
       "                                                                                \n",
       "       <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                       </span>       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['mlp', 'cnn']                       </span>       \n",
       "        lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.1, End: 0.5, Prior: uniform </span>       \n",
       "        batch_size    integer      <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 1, End: 5, Prior: log-uniform </span>       \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "      \u001b[3m                 MLE-Hyperopt Halving Hyperspace 🚀                 \u001b[0m      \n",
       "                                                                                \n",
       "       \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                      \u001b[0m\u001b[1;31m \u001b[0m       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  \u001b[31m \u001b[0m\u001b[31m['mlp', 'cnn']                      \u001b[0m\u001b[31m \u001b[0m       \n",
       "        lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.1, End: 0.5, Prior: uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "        batch_size    integer      \u001b[31m \u001b[0m\u001b[31mBegin: 1, End: 5, Prior: log-uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:07] </span>Start running <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> batches of Successive Halving.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#186\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">186</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:07]\u001b[0m\u001b[2;36m \u001b[0mStart running \u001b[1;36m5\u001b[0m batches of Successive Halving.           \u001b]8;id=812109;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=680980;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#186\u001b\\\u001b[2m186\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Configurations per batch: <span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">20</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">10</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span><span style=\"font-weight: bold\">]</span>            <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#189\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">189</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Configurations per batch: \u001b[1m[\u001b[0m\u001b[1;36m20\u001b[0m, \u001b[1;36m10\u001b[0m, \u001b[1;36m5\u001b[0m, \u001b[1;36m2\u001b[0m, \u001b[1;36m1\u001b[0m\u001b[1m]\u001b[0m            \u001b]8;id=707594;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=673744;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#189\u001b\\\u001b[2m189\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Iterations per batch: <span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">8</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">16</span><span style=\"font-weight: bold\">]</span>                 <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#190\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">190</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Iterations per batch: \u001b[1m[\u001b[0m\u001b[1;36m1\u001b[0m, \u001b[1;36m2\u001b[0m, \u001b[1;36m4\u001b[0m, \u001b[1;36m8\u001b[0m, \u001b[1;36m16\u001b[0m\u001b[1m]\u001b[0m                 \u001b]8;id=809262;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=769056;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#190\u001b\\\u001b[2m190\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Halving coefficient: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>                                 <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#191\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">191</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Halving coefficient: \u001b[1;36m2\u001b[0m                                 \u001b]8;id=925795;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=106955;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#191\u001b\\\u001b[2m191\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Total Number of Iterations: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">92</span>                         <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#192\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">192</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Total Number of Iterations: \u001b[1;36m92\u001b[0m                         \u001b]8;id=225188;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=774092;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#192\u001b\\\u001b[2m192\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Batch No. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">20</span> configs for <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> iters.                 <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#193\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">193</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Batch No. \u001b[1;36m1\u001b[0m/\u001b[1;36m5\u001b[0m: \u001b[1;36m20\u001b[0m configs for \u001b[1;36m1\u001b[0m iters.                 \u001b]8;id=736925;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=59646;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#193\u001b\\\u001b[2m193\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mle_hyperopt import HalvingSearch\n",
    "\n",
    "strategy = HalvingSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                               \"end\": 0.5,\n",
    "                               \"prior\": \"uniform\"}},\n",
    "                         integer={\"batch_size\": {\"begin\": 1,\n",
    "                                                 \"end\": 5,\n",
    "                                                 \"prior\": \"log-uniform\"}},\n",
    "                         categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                         search_config={\"min_budget\": 1,\n",
    "                                        \"num_arms\": 20,\n",
    "                                        \"halving_coeff\": 2},\n",
    "                         seed_id=42,\n",
    "                         verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "e3331c6b-39e8-4520-9f09-46fdf33a6a3e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 20  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:23        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 13 │ 10.185  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_0_13.pt'                     │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 13 │ 10.185  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_0_13.pt'                     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 20 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:23       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 13 │ 10.185  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_0_13.pt'                     │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 13 │ 10.185  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_0_13.pt'                     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:23] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> batches of SH ➢ <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">20</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">92</span> iters.               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#221\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">221</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:23]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m1\u001b[0m/\u001b[1;36m5\u001b[0m batches of SH ➢ \u001b[1;36m20\u001b[0m/\u001b[1;36m92\u001b[0m iters.               \u001b]8;id=450320;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=207034;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#221\u001b\\\u001b[2m221\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Next - Batch No. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">10</span> configs for <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> iters.          <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#226\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">226</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Next - Batch No. \u001b[1;36m2\u001b[0m/\u001b[1;36m5\u001b[0m: \u001b[1;36m10\u001b[0m configs for \u001b[1;36m2\u001b[0m iters.          \u001b]8;id=258589;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=702177;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#226\u001b\\\u001b[2m226\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "configs = strategy.ask()\n",
    "scores = [get_iteration_score(c[\"extra\"][\"sh_num_total_iters\"], 0, **c[\"params\"])[1] for c in configs]\n",
    "ckpts = [f\"ckpt_0_{i}.pt\" for i in range(len(configs))]\n",
    "strategy.tell(configs, scores, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "246d95f4-1f28-45cc-bc54-aeea0216fc15",
   "metadata": {},
   "source": [
    "We want to be able to keep track of the previously trained checkpoints. If a configuration \"survives\" we can amortize the previous training steps by reloading the old checkpoint and only train for the remaining additional step. The corresponding configurations therefore have a different structure, which allows to access the checkpoint via `extra`-> `sh_ckpt`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "4764b9c3-9de4-4108-b66a-299ca413c9e4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'params': {'arch': 'cnn', 'lrate': 0.48, 'batch_size': 4}, 'extra': {'sh_num_total_iters': 2, 'sh_num_add_iters': 1, 'sh_counter': 1, 'sh_ckpt': 'ckpt_0_13.pt'}}\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 30  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:25        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 20 │ 5.155   │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_1_0.pt'                      │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 20 │ 5.155   │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_1_0.pt'                      │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 30 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:25       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 20 │ 5.155   │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_1_0.pt'                      │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 20 │ 5.155   │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_1_0.pt'                      │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:25] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> batches of SH ➢ <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">40</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">92</span> iters.               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#221\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">221</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:25]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m2\u001b[0m/\u001b[1;36m5\u001b[0m batches of SH ➢ \u001b[1;36m40\u001b[0m/\u001b[1;36m92\u001b[0m iters.               \u001b]8;id=504502;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=311393;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#221\u001b\\\u001b[2m221\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Next - Batch No. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> configs for <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> iters.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#226\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">226</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Next - Batch No. \u001b[1;36m3\u001b[0m/\u001b[1;36m5\u001b[0m: \u001b[1;36m5\u001b[0m configs for \u001b[1;36m4\u001b[0m iters.           \u001b]8;id=377179;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=115938;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#226\u001b\\\u001b[2m226\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "configs = strategy.ask()\n",
    "print(configs[0])\n",
    "scores = [get_iteration_score(c[\"extra\"][\"sh_num_total_iters\"], 0, **c[\"params\"])[1] for c in configs]\n",
    "ckpts = [f\"ckpt_1_{i}.pt\" for i in range(len(configs))]\n",
    "strategy.tell(configs, scores, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3519d748-b798-4a6c-976e-45d2772f0e97",
   "metadata": {},
   "source": [
    "Another point worth noting is that you can arbitrarily redefine the amount of GD steps which equal one SH iteration for your application at hand. For such a more realistic example which trains a small MLP, I recommend to check out this [example script](https://github.com/mle-infrastructure/mle-hyperopt/blob/main/examples/run_sh_mlp.py)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf3131d2-73b0-4c92-8dc1-cbadcb0f5613",
   "metadata": {},
   "source": [
    "## Hyperband <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/hyperband.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "Hyperband [(Li et al., 2018)](https://arxiv.org/pdf/1603.06560.pdf) provides an extension to Successive Halving. In SH it is not apriori clear whether we would like to evaluate a small set of configurations for many iterations or instead explore many of them for only a few iterations. You as a designer have to tune the trade-off. Hyperband tries to overcome this problem by calling SH as an inner loop for different trade-off settings. Simply put: Hyperband runs many SH loops with different hyper-hyperparameters. You only have to set the maximal number of iterations that can be prescribed to any single run (`max_resource`) and the having aggressivity (`eta`). That's all.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "c0d4c9e6-83f6-46cb-9f87-abd40a247d63",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">      <span style=\"font-style: italic\">                MLE-Hyperopt Hyperband Hyperspace 🚀                </span>      \n",
       "                                                                                \n",
       "       <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type        </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                       </span>       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  <span style=\"color: #800000; text-decoration-color: #800000\"> ['mlp', 'cnn']                       </span>       \n",
       "        lrate         real         <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.1, End: 0.5, Prior: uniform </span>       \n",
       "        batch_size    integer      <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 1, End: 5, Prior: log-uniform </span>       \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "      \u001b[3m                MLE-Hyperopt Hyperband Hyperspace 🚀                \u001b[0m      \n",
       "                                                                                \n",
       "       \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType       \u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                      \u001b[0m\u001b[1;31m \u001b[0m       \n",
       "       ──────────────────────────────────────────────────────────────────       \n",
       "        arch          categorical  \u001b[31m \u001b[0m\u001b[31m['mlp', 'cnn']                      \u001b[0m\u001b[31m \u001b[0m       \n",
       "        lrate         real         \u001b[31m \u001b[0m\u001b[31mBegin: 0.1, End: 0.5, Prior: uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "        batch_size    integer      \u001b[31m \u001b[0m\u001b[31mBegin: 1, End: 5, Prior: log-uniform\u001b[0m\u001b[31m \u001b[0m       \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:27] </span>Start running <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">11</span> batches of Hyperband evaluations.       <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#250\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">250</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:27]\u001b[0m\u001b[2;36m \u001b[0mStart running \u001b[1;36m11\u001b[0m batches of Hyperband evaluations.       \u001b]8;id=790308;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=906044;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#250\u001b\\\u001b[2m250\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Evals per batch: <span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">27</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">9</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">9</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span><span style=\"font-weight: bold\">]</span>    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#253\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">253</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Evals per batch: \u001b[1m[\u001b[0m\u001b[1;36m27\u001b[0m, \u001b[1;36m9\u001b[0m, \u001b[1;36m3\u001b[0m, \u001b[1;36m1\u001b[0m, \u001b[1;36m9\u001b[0m, \u001b[1;36m3\u001b[0m, \u001b[1;36m1\u001b[0m, \u001b[1;36m6\u001b[0m, \u001b[1;36m2\u001b[0m, \u001b[1;36m4\u001b[0m, \u001b[1;36m1\u001b[0m\u001b[1m]\u001b[0m    \u001b]8;id=732274;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=843156;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#253\u001b\\\u001b[2m253\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Total SH loops: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> | Arms per loop: <span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">27</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">9</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span><span style=\"font-weight: bold\">]</span>       <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#254\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">254</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Total SH loops: \u001b[1;36m4\u001b[0m | Arms per loop: \u001b[1m[\u001b[0m\u001b[1;36m27\u001b[0m, \u001b[1;36m9\u001b[0m, \u001b[1;36m6\u001b[0m, \u001b[1;36m4\u001b[0m\u001b[1m]\u001b[0m       \u001b]8;id=733989;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=214082;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#254\u001b\\\u001b[2m254\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Min. budget per loop: <span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">9</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">27</span><span style=\"font-weight: bold\">]</span>                    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#257\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">257</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Min. budget per loop: \u001b[1m[\u001b[0m\u001b[1;36m1\u001b[0m, \u001b[1;36m3\u001b[0m, \u001b[1;36m9\u001b[0m, \u001b[1;36m27\u001b[0m\u001b[1m]\u001b[0m                    \u001b]8;id=994303;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=826236;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#257\u001b\\\u001b[2m257\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Start Loop No. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">27</span> arms &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> min budget.            <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#258\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">258</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Start Loop No. \u001b[1;36m1\u001b[0m/\u001b[1;36m4\u001b[0m: \u001b[1;36m27\u001b[0m arms & \u001b[1;36m1\u001b[0m min budget.            \u001b]8;id=137769;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=3243;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#258\u001b\\\u001b[2m258\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mle_hyperopt import HyperbandSearch\n",
    "\n",
    "strategy = HyperbandSearch(real={\"lrate\": {\"begin\": 0.1,\n",
    "                                        \"end\": 0.5,\n",
    "                                        \"prior\": \"uniform\"}},\n",
    "                           integer={\"batch_size\": {\"begin\": 1,\n",
    "                                                   \"end\": 5,\n",
    "                                                   \"prior\": \"log-uniform\"}},\n",
    "                           categorical={\"arch\": [\"mlp\", \"cnn\"]},\n",
    "                           search_config={\"max_resource\": 27,\n",
    "                                          \"eta\": 3},\n",
    "                           seed_id=42,\n",
    "                           verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "83da65f0-995a-4c71-9772-72b55b7fa63f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 27  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:27        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 13 │ 10.333  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_2_13.pt'                     │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 13 │ 10.333  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ 4, 'ckpt': 'ckpt_2_13.pt'                     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 27 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:27       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 13 │ 10.333  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_2_13.pt'                     │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 13 │ 10.333  │ 'arch': 'cnn', 'lrate': 0.48, 'batch_size':   │\n",
       "│\u001b[2m               \u001b[0m│    │         │ 4, 'ckpt': 'ckpt_2_13.pt'                     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:27] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">11</span> of Hyperband evaluation batches.          <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#285\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">285</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:27]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m1\u001b[0m/\u001b[1;36m11\u001b[0m of Hyperband evaluation batches.          \u001b]8;id=619858;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=371824;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#285\u001b\\\u001b[2m285\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Done with <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> loops of SH.                             <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#289\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">289</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Done with \u001b[1;36m0\u001b[0m/\u001b[1;36m4\u001b[0m loops of SH.                             \u001b]8;id=103073;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=487971;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#289\u001b\\\u001b[2m289\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Active Loop No. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>/<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">27</span> arms &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> min budget.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#291\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">291</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Active Loop No. \u001b[1;36m1\u001b[0m/\u001b[1;36m4\u001b[0m: \u001b[1;36m27\u001b[0m arms & \u001b[1;36m1\u001b[0m min budget.           \u001b]8;id=716260;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=86448;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#291\u001b\\\u001b[2m291\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Next batch of SH: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">9</span> evals.                             <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#296\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">296</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Next batch of SH: \u001b[1;36m9\u001b[0m evals.                             \u001b]8;id=795379;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=868298;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#296\u001b\\\u001b[2m296\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "configs = strategy.ask()\n",
    "scores = [get_iteration_score(c[\"extra\"][\"sh_num_total_iters\"], 0, **c[\"params\"])[1] for c in configs]\n",
    "ckpts = [f\"ckpt_2_{i}.pt\" for i in range(len(configs))]\n",
    "strategy.tell(configs, scores, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b65efe0e-10ea-4311-8773-a031cfa4c946",
   "metadata": {},
   "source": [
    "## Population-Based Training <img src=\"https://github.com/mle-infrastructure/mle-hyperopt/blob/main/docs/pbt.png?raw=true\" alt=\"drawing\" width=\"50\"/>\n",
    "\n",
    "`mle-hyperopt` also implements Population-Based Training [(Jaderberg et al., 2017)](https://arxiv.org/pdf/1711.09846.pdf). In the following we replicate the toy example of figure 2 in the main paper. You can find a neural network example [here](https://github.com/mle-infrastructure/mle-hyperopt/blob/main/examples/run_pbt_mlp.py) with PyTorch training loop."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "20d192dd-2741-484a-88e4-9ca58988466e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from mle_hyperopt import PBTSearch\n",
    "\n",
    "class QuadraticProblem(object):\n",
    "    def __init__(self, theta=None, lrate: float = 0.1):\n",
    "        self.lrate = lrate\n",
    "\n",
    "    def step(self, h, theta):\n",
    "        \"\"\"Perform GradAscent step on quadratic surrogate objective (max!).\"\"\"\n",
    "        surrogate_grad = -2.0 * h * theta\n",
    "        return theta + self.lrate * surrogate_grad\n",
    "\n",
    "    def evaluate(self, theta):\n",
    "        \"\"\"Ground truth objective (e.g. val loss) - Jaderberg et al. 2016.\"\"\"\n",
    "        return 1.2 - np.sum(theta ** 2)\n",
    "\n",
    "    def surrogate_objective(self, h, theta):\n",
    "        \"\"\"Surrogate objective (with hyperparams h) - Jaderberg et al. 2016.\"\"\"\n",
    "        return 1.2 - np.sum(h * theta ** 2)\n",
    "\n",
    "    def __call__(self, theta, hyperparams):\n",
    "        h = np.array([hyperparams[\"h0\"], hyperparams[\"h1\"]])\n",
    "        theta = self.step(h, theta)\n",
    "        exact = self.evaluate(theta)\n",
    "        surrogate = self.surrogate_objective(h, theta)\n",
    "        return theta.tolist(), exact        \n",
    "\n",
    "\n",
    "problem = QuadraticProblem()\n",
    "ckpts = [[0.9, 0.9], [0.9, 0.9]]\n",
    "value = [problem(ckpts[0], {\"h0\": 0.9, \"h1\": 0.9})[1],\n",
    "         problem(ckpts[1], {\"h0\": 0.9, \"h1\": 0.9})[1]]\n",
    "\n",
    "theta_log = [ckpts]\n",
    "value_log = [value]\n",
    "steps_until_ready = 4"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93fd08a5-d5c7-4813-a7ad-9f6bc4684713",
   "metadata": {},
   "source": [
    "The exploration and exploitation strategies can be specified using the `explore`, `exploit` settings in the `search_config`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "84b2ed77-61ba-4299-b429-87a4bb421f12",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">         <span style=\"font-style: italic\">               MLE-Hyperopt PBT Hyperspace 🚀                </span>          \n",
       "                                                                                \n",
       "          <span style=\"font-weight: bold\"> 🌻 Variable </span> <span style=\"font-weight: bold\"> Type </span> <span style=\"color: #800000; text-decoration-color: #800000; font-weight: bold\"> Search Range ↔                       </span>           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           h0            real  <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.0, End: 1.0, Prior: uniform </span>           \n",
       "           h1            real  <span style=\"color: #800000; text-decoration-color: #800000\"> Begin: 0.0, End: 1.0, Prior: uniform </span>           \n",
       "                                                                                \n",
       "</pre>\n"
      ],
      "text/plain": [
       "         \u001b[3m               MLE-Hyperopt PBT Hyperspace 🚀                \u001b[0m          \n",
       "                                                                                \n",
       "          \u001b[1m \u001b[0m\u001b[1m🌻 Variable\u001b[0m\u001b[1m \u001b[0m \u001b[1m \u001b[0m\u001b[1mType\u001b[0m\u001b[1m \u001b[0m \u001b[1;31m \u001b[0m\u001b[1;31mSearch Range ↔                      \u001b[0m\u001b[1;31m \u001b[0m           \n",
       "          ───────────────────────────────────────────────────────────           \n",
       "           h0            real  \u001b[31m \u001b[0m\u001b[31mBegin: 0.0, End: 1.0, Prior: uniform\u001b[0m\u001b[31m \u001b[0m           \n",
       "           h1            real  \u001b[31m \u001b[0m\u001b[31mBegin: 0.0, End: 1.0, Prior: uniform\u001b[0m\u001b[31m \u001b[0m           \n",
       "                                                                                \n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:32] </span>Start running PBT w. <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> workers.                          <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#316\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">316</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:32]\u001b[0m\u001b[2;36m \u001b[0mStart running PBT w. \u001b[1;36m2\u001b[0m workers.                          \u001b]8;id=795101;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=617236;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#316\u001b\\\u001b[2m316\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Steps until ready: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>                                   <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#317\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">317</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Steps until ready: \u001b[1;36m4\u001b[0m                                   \u001b]8;id=757733;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=314991;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#317\u001b\\\u001b[2m317\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Exploration strategy: additive-noise                   <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#318\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">318</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Exploration strategy: additive-noise                   \u001b]8;id=717287;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=769132;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#318\u001b\\\u001b[2m318\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Exploitation strategy: truncation                      <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#319\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">319</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Exploitation strategy: truncation                      \u001b]8;id=988681;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=524014;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#319\u001b\\\u001b[2m319\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "strategy = PBTSearch(\n",
    "    real={\n",
    "        \"h0\": {\"begin\": 0.0, \"end\": 1.0, \"prior\": \"uniform\"},\n",
    "        \"h1\": {\"begin\": 0.0, \"end\": 1.0, \"prior\": \"uniform\"},\n",
    "    },\n",
    "    search_config={\n",
    "        \"exploit\": {\"strategy\": \"truncation\", \"selection_percent\": 0.2},\n",
    "        \"explore\": {\"strategy\": \"additive-noise\", \"noise_scale\": 0.35},\n",
    "        \"steps_until_ready\": 4,\n",
    "        \"num_workers\": 2,\n",
    "    },\n",
    "    maximize_objective=True,\n",
    "    verbose=True\n",
    ")\n",
    "\n",
    "configs = strategy.ask()\n",
    "configs[0]['params']['h0'] = 0.1\n",
    "configs[0]['params']['h1'] = 0.9\n",
    "configs[1]['params']['h0'] = 0.9\n",
    "configs[1]['params']['h1'] = 0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "0753153f-bd3f-4a3f-ae78-ed49413fd396",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 2   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:33        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 1  │ 0.345   │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.40690958399999994, 0.8301313440000001]     │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 0  │ 0.345   │ 'h0': 0.1, 'h1': 0.9, 'ckpt':                 │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.8301313440000001, 0.40690958399999994]     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 2  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:33       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 1  │ 0.345   │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.40690958399999994, 0.8301313440000001]     │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 0  │ 0.345   │ 'h0': 0.1, 'h1': 0.9, 'ckpt':                 │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.8301313440000001, 0.40690958399999994]     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:33] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:33]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m1\u001b[0m batches of PBT.                              \u001b]8;id=485348;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=486140;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>                               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m4\u001b[0m                               \u001b]8;id=80456;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=10504;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.345</span><span style=\"font-weight: bold\">)</span> exploits W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.345</span><span style=\"font-weight: bold\">)</span>              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m0.345\u001b[0m\u001b[1m)\u001b[0m exploits W1 \u001b[1m(\u001b[0mP: \u001b[1;36m0.345\u001b[0m\u001b[1m)\u001b[0m              \u001b]8;id=792641;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=845373;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.9</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.1</span><span style=\"font-weight: bold\">}</span>                    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m0.9\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.1\u001b[0m\u001b[1m}\u001b[0m                    \u001b]8;id=136607;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=476349;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.345</span><span style=\"font-weight: bold\">)</span> continues own trajectory.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m0.345\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.           \u001b]8;id=817757;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=658658;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.9</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.1</span><span style=\"font-weight: bold\">}</span>                    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m0.9\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.1\u001b[0m\u001b[1m}\u001b[0m                    \u001b]8;id=972354;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=549072;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for s in range(steps_until_ready):\n",
    "    theta_l, value_l = [], []\n",
    "    for i in range(len(configs)):\n",
    "        theta, exact = problem(theta_log[-1][i], configs[i][\"params\"])\n",
    "        theta_l.append(theta)\n",
    "        value_l.append(exact)\n",
    "    value_log.append(value_l)\n",
    "    theta_log.append(theta_l)\n",
    "\n",
    "ckpts = theta_log[-1]\n",
    "values = value_log[-1]\n",
    "strategy.tell(configs, values, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fda8fca-9c91-41d1-92b2-598b8ee5f76a",
   "metadata": {},
   "source": [
    "Next we ask for the next set of PBT hyperparameters and the checkpoints to reload. Note that in our simple example we directly use the `theta` values as `pbt_ckpt`. In more realistic settings this will be a string that points you to the network checkpoint to reload. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "cf897f6c-9e0a-4fd8-9560-f69084e44fc6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 4   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:36        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 3  │ 0.58    │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.1839726772789478, 0.7656867203236071]      │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 3  │ 0.58    │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.1839726772789478, 0.7656867203236071]      │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 4  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:36       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 3  │ 0.58    │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.1839726772789478, 0.7656867203236071]      │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 3  │ 0.58    │ 'h0': 0.9, 'h1': 0.1, 'ckpt':                 │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.1839726772789478, 0.7656867203236071]      │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:36] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:36]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m2\u001b[0m batches of PBT.                              \u001b]8;id=996250;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=534488;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">8</span>                               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m8\u001b[0m                               \u001b]8;id=408194;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=49160;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.491</span><span style=\"font-weight: bold\">)</span> exploits W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.58</span><span style=\"font-weight: bold\">)</span>               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m0.491\u001b[0m\u001b[1m)\u001b[0m exploits W1 \u001b[1m(\u001b[0mP: \u001b[1;36m0.58\u001b[0m\u001b[1m)\u001b[0m               \u001b]8;id=33803;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=886604;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.9</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.1</span><span style=\"font-weight: bold\">}</span>                    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m0.9\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.1\u001b[0m\u001b[1m}\u001b[0m                    \u001b]8;id=542870;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=55856;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.58</span><span style=\"font-weight: bold\">)</span> continues own trajectory.            <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m0.58\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.            \u001b]8;id=533984;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=142271;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.9</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.1</span><span style=\"font-weight: bold\">}</span>                    <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m0.9\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.1\u001b[0m\u001b[1m}\u001b[0m                    \u001b]8;id=679483;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=939615;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 6   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:36        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 4  │ 0.88    │ 'h0': 1.453, 'h1': 0.369, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.04660719817215214, 0.5636629049030013]     │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 4  │ 0.88    │ 'h0': 1.453, 'h1': 0.369, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.04660719817215214, 0.5636629049030013]     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 6  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:36       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 4  │ 0.88    │ 'h0': 1.453, 'h1': 0.369, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.04660719817215214, 0.5636629049030013]     │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 4  │ 0.88    │ 'h0': 1.453, 'h1': 0.369, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.04660719817215214, 0.5636629049030013]     │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:36] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:36]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m3\u001b[0m batches of PBT.                              \u001b]8;id=536447;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=950952;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">12</span>                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m12\u001b[0m                              \u001b]8;id=843592;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=419406;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.88</span><span style=\"font-weight: bold\">)</span> continues own trajectory.            <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m0.88\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.            \u001b]8;id=354864;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=860645;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.453</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.369</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.453\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.369\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=24094;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=380900;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.694</span><span style=\"font-weight: bold\">)</span> exploits W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.88</span><span style=\"font-weight: bold\">)</span>               <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m0.694\u001b[0m\u001b[1m)\u001b[0m exploits W0 \u001b[1m(\u001b[0mP: \u001b[1;36m0.88\u001b[0m\u001b[1m)\u001b[0m               \u001b]8;id=169271;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=841960;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.453</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.369</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.453\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.369\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=445304;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=134783;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 8   </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:36        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 7  │ 1.077   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.014151865342057996, 0.35096096741021804]   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 7  │ 1.077   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.014151865342057996, 0.35096096741021804]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 8  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:36       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 7  │ 1.077   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.014151865342057996, 0.35096096741021804]   │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 7  │ 1.077   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.014151865342057996, 0.35096096741021804]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:36] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:36]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m4\u001b[0m batches of PBT.                              \u001b]8;id=611578;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=941975;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">16</span>                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m16\u001b[0m                              \u001b]8;id=330472;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=348688;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.028</span><span style=\"font-weight: bold\">)</span> exploits W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.077</span><span style=\"font-weight: bold\">)</span>              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m1.028\u001b[0m\u001b[1m)\u001b[0m exploits W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.077\u001b[0m\u001b[1m)\u001b[0m              \u001b]8;id=602455;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=867852;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=424567;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=274332;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.077</span><span style=\"font-weight: bold\">)</span> continues own trajectory.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.077\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.           \u001b]8;id=900960;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=412664;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=658846;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=435078;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 10  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:36        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 9  │ 1.152   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.004297089301956946, 0.21852351746778978]   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 9  │ 1.152   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.004297089301956946, 0.21852351746778978]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 10 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:36       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 9  │ 1.152   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.004297089301956946, 0.21852351746778978]   │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 9  │ 1.152   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.004297089301956946, 0.21852351746778978]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:36] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:36]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m5\u001b[0m batches of PBT.                              \u001b]8;id=875588;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=185092;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">20</span>                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m20\u001b[0m                              \u001b]8;id=668907;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=641299;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.136</span><span style=\"font-weight: bold\">)</span> exploits W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.152</span><span style=\"font-weight: bold\">)</span>              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m1.136\u001b[0m\u001b[1m)\u001b[0m exploits W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.152\u001b[0m\u001b[1m)\u001b[0m              \u001b]8;id=802405;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=579237;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=9616;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=256299;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.152</span><span style=\"font-weight: bold\">)</span> continues own trajectory.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.152\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.           \u001b]8;id=499242;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=919415;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=54641;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=947882;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> 📥 Total: 12  </span>┃<span style=\"font-weight: bold\"> ID </span>┃<span style=\"font-weight: bold\"> Obj. 📉 </span>┃<span style=\"font-weight: bold\"> Configuration 🔖 - 03/08/2023 13:01:36        </span>┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best Overall  </span>│ 11 │ 1.181   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.0013047733300653084, 0.1360622180833011]   │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\"> Best in Batch </span>│ 11 │ 1.181   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│<span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">               </span>│    │         │ [0.0013047733300653084, 0.1360622180833011]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━┳━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1m📥 Total: 12 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mID\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mObj. 📉\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConfiguration 🔖 - 03/08/2023 13:01:36       \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━╇━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest Overall \u001b[0m\u001b[2m \u001b[0m│ 11 │ 1.181   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.0013047733300653084, 0.1360622180833011]   │\n",
       "│\u001b[2m \u001b[0m\u001b[2mBest in Batch\u001b[0m\u001b[2m \u001b[0m│ 11 │ 1.181   │ 'h0': 1.288, 'h1': 0.558, 'ckpt':             │\n",
       "│\u001b[2m               \u001b[0m│    │         │ [0.0013047733300653084, 0.1360622180833011]   │\n",
       "└───────────────┴────┴─────────┴───────────────────────────────────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">[13:01:36] </span>Completed <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span> batches of PBT.                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">333</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m[13:01:36]\u001b[0m\u001b[2;36m \u001b[0mCompleted \u001b[1;36m6\u001b[0m batches of PBT.                              \u001b]8;id=681084;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=25328;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#333\u001b\\\u001b[2m333\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ Number of total steps: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">24</span>                              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">334</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ Number of total steps: \u001b[1;36m24\u001b[0m                              \u001b]8;id=923002;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=671174;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#334\u001b\\\u001b[2m334\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W0 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.143</span><span style=\"font-weight: bold\">)</span> exploits W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.181</span><span style=\"font-weight: bold\">)</span>              <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">337</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W0 \u001b[1m(\u001b[0mP: \u001b[1;36m1.143\u001b[0m\u001b[1m)\u001b[0m exploits W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.181\u001b[0m\u001b[1m)\u001b[0m              \u001b]8;id=856725;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=962399;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#337\u001b\\\u001b[2m337\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- E/E Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">343</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- E/E Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=653489;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=163077;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#343\u001b\\\u001b[2m343\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>➞ 👨‍🚒 W1 <span style=\"font-weight: bold\">(</span>P: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.181</span><span style=\"font-weight: bold\">)</span> continues own trajectory.           <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">345</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m➞ 👨‍🚒 W1 \u001b[1m(\u001b[0mP: \u001b[1;36m1.181\u001b[0m\u001b[1m)\u001b[0m continues own trajectory.           \u001b]8;id=694003;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=260808;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#345\u001b\\\u001b[2m345\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #7fbfbf; text-decoration-color: #7fbfbf\">           </span>-- Old Params: <span style=\"font-weight: bold\">{</span><span style=\"color: #008000; text-decoration-color: #008000\">'h0'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1.288</span>, <span style=\"color: #008000; text-decoration-color: #008000\">'h1'</span>: <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.558</span><span style=\"font-weight: bold\">}</span>                <a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">comms.py</span></a><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">:</span><a href=\"file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\"><span style=\"color: #7f7f7f; text-decoration-color: #7f7f7f\">350</span></a>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[2;36m          \u001b[0m\u001b[2;36m \u001b[0m-- Old Params: \u001b[1m{\u001b[0m\u001b[32m'h0'\u001b[0m: \u001b[1;36m1.288\u001b[0m, \u001b[32m'h1'\u001b[0m: \u001b[1;36m0.558\u001b[0m\u001b[1m}\u001b[0m                \u001b]8;id=964062;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py\u001b\\\u001b[2mcomms.py\u001b[0m\u001b]8;;\u001b\\\u001b[2m:\u001b[0m\u001b]8;id=791631;file:///Users/rob/Dropbox/core-code/mle-infrastructure/mle-hyperopt/mle_hyperopt/utils/comms.py#350\u001b\\\u001b[2m350\u001b[0m\u001b]8;;\u001b\\\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_pbt_steps = 5\n",
    "for t in range(num_pbt_steps):\n",
    "    configs = strategy.ask()\n",
    "    for s in range(steps_until_ready):\n",
    "        theta_l, value_l = [], []\n",
    "        for i in range(len(configs)):\n",
    "            if s == 0:\n",
    "                theta_log[-1][i] = configs[i][\"extra\"][\"pbt_ckpt\"]\n",
    "            theta, exact = problem(theta_log[-1][i], configs[i][\"params\"])\n",
    "            theta_l.append(theta)\n",
    "            value_l.append(exact)\n",
    "        value_log.append(value_l)\n",
    "        theta_log.append(theta_l)\n",
    "\n",
    "    ckpts = theta_log[-1]\n",
    "    values = value_log[-1]\n",
    "    strategy.tell(configs, values, ckpts)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d34856f1-4f2d-4c2c-b8a6-ee19b89abdd1",
   "metadata": {},
   "source": [
    "**Note**: The `theta` plot below is different from the one shown in the paper (figure 2). There are two potential reasons for this: \n",
    "\n",
    "1. We use a synchronous version of PBT, which waits until all workers have completed their 4 gradient descent steps. The original version is asynchronous and each worker writes its state to memory/a file system after each individual evaluation. Thereby slow workers can be pulled ahead by copying fast workers.\n",
    "2. The difference between the sync and async PBT setting are more likely to have impact when there are many workers. The here observed differences are more likely due to the fact that the toy problem in Jaderberg et al. (2017) actually does not implement the proposed copying of the `h` hyperparameters before applying the exploration. Instead each worker only copies the `theta` state and keeps its own `h`. Afterwards, these are slightly perturbed. I assume the paper chose to do so as an educational/illustrative excercise, given that copying also `h` increases performance! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "b5a0f7f4-6400-404a-8c01-ba57cecf000c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 682,
       "width": 664
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "fig, axs = plt.subplots(2, 2, figsize=(10, 10))\n",
    "axs.flatten()[0].scatter(np.array(theta_log)[:, 0, 1], np.array(theta_log)[:, 0, 0], c=\"b\", label=\"W-1\", alpha=0.5)\n",
    "axs.flatten()[0].scatter(np.array(theta_log)[:, 1, 1], np.array(theta_log)[:, 1, 0], c=\"r\", label=\"W-2\", alpha=0.5)\n",
    "axs.flatten()[0].set_xlim(-0.1, 1.1)\n",
    "axs.flatten()[0].set_ylim(-0.1, 1.1)\n",
    "axs.flatten()[0].set_xlabel(r\"$\\theta_0$\")\n",
    "axs.flatten()[0].set_ylabel(r\"$\\theta_1$\")\n",
    "axs.flatten()[0].set_title(\"Optim. Params\")\n",
    "\n",
    "axs.flatten()[1].plot(np.array(value_log)[:, 0], c=\"b\", label=\"Worker 1\", alpha=0.5)\n",
    "axs.flatten()[1].plot(np.array(value_log)[:, 1], c=\"r\", label=\"Worker 2\", alpha=0.5)\n",
    "axs.flatten()[1].axhline(1.2, c=\"y\", ls=\"--\", label=\"Optimum\")\n",
    "axs.flatten()[1].set_xlabel(\"# GD Steps\")\n",
    "axs.flatten()[1].set_ylabel(r\"$Q$ - Max. Target\")\n",
    "axs.flatten()[1].legend(fontsize=20)\n",
    "axs.flatten()[1].set_title(\"Worker Fitness\")\n",
    "\n",
    "axs.flatten()[2].plot(strategy.df[strategy.df['pbt_worker_id']==0].h0.values,\n",
    "                      c=\"b\", label=\"W-1\", alpha=0.5)\n",
    "axs.flatten()[2].plot(strategy.df[strategy.df['pbt_worker_id']==1].h0.values,\n",
    "                      c=\"r\", label=\"W-2\", alpha=0.5)\n",
    "axs.flatten()[2].set_xlabel(r\"# PBT Steps\")\n",
    "axs.flatten()[2].set_ylabel(r\"$h_0$\")\n",
    "axs.flatten()[2].set_title(r\"Hyper: $h_0$\")\n",
    "\n",
    "axs.flatten()[3].plot(strategy.df[strategy.df['pbt_worker_id']==0].h1.values,\n",
    "                      c=\"b\", label=\"W-1\", alpha=0.5)\n",
    "axs.flatten()[3].plot(strategy.df[strategy.df['pbt_worker_id']==1].h1.values,\n",
    "                      c=\"r\", label=\"W-2\", alpha=0.5)\n",
    "axs.flatten()[3].set_xlabel(r\"# PBT Steps\")\n",
    "axs.flatten()[3].set_ylabel(r\"$h_1$\")\n",
    "axs.flatten()[3].set_title(r\"Hyper: $h_1$\")\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93e59b79-af74-44f1-be15-c874dd4e1577",
   "metadata": {},
   "source": [
    "That's it! Give `mle-hyperopt` a try and let me know what you think! If you find a bug or are missing your favourite feature, feel free to contact me [@RobertTLange](https://twitter.com/RobertTLange) or create an issue!"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "0731916ac994366e3ec5c9b7f6ac08b5be0d09771460287bbfce9f645047acc4"
  },
  "kernelspec": {
   "display_name": "mle-toolbox",
   "language": "python",
   "name": "mle-toolbox"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  },
  "toc-showcode": false
 },
 "nbformat": 4,
 "nbformat_minor": 5
}