framed_hopper.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "from pathlib import Path\n",
    "import functools\n",
    "\n",
    "import mujoco\n",
    "from brax import envs\n",
    "import jax\n",
    "import jax.numpy as jp\n",
    "\n",
    "from envs import register_framed_hopper\n",
    "from shac.brax_wrappers import AppendObsHistory\n",
    "\n",
    "import shac.networks as shac_networks\n",
    "# import shac.train as  shac\n",
    "from shac.train import SHAC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "env_name = 'framed_hopper'\n",
    "\n",
    "log_sigma = None\n",
    "make_networks_factory = functools.partial(\n",
    "    shac_networks.make_shac_networks,\n",
    "        policy_hidden_layer_sizes=(64, 64),\n",
    "        value_hidden_layer_sizes =(64, 64),\n",
    "        scalar_var=False,\n",
    "        layer_norm=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Parameters\n",
    "num_epochs = 200\n",
    "num_evals = num_epochs + 1\n",
    "\n",
    "unroll_length = 32\n",
    "\n",
    "num_envs = 64\n",
    "\n",
    "num_timesteps = (num_envs*unroll_length)*num_epochs\n",
    "\n",
    "num_critic_minibatches = 4\n",
    "num_critic_data = num_envs*unroll_length\n",
    "critic_batch_size = num_critic_data / num_critic_minibatches\n",
    "\n",
    "episode_length = 240 \n",
    "\n",
    "# Logging\n",
    "num_grad_checks = None\n",
    "save_all_policy_gradients = True\n",
    "\n",
    "# Env parameters\n",
    "action_pad = 1e-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "env      = AppendObsHistory(envs.get_environment(env_name, action_pad=action_pad), h=3)\n",
    "eval_env = AppendObsHistory(envs.get_environment(env_name, action_pad=action_pad), h=3)\n",
    "# env = envs.get_environment(env_name)\n",
    "# eval_env = envs.get_environment(env_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Env steps per training step: 2048\n",
      "Training steps per epoch: 1\n",
      "Critic minibatches per critic epoch: 4\n"
     ]
    }
   ],
   "source": [
    "kwargs = dict(\n",
    "  num_timesteps=num_timesteps, log_sigma=log_sigma,\n",
    "  num_evals=num_evals, use_tbx=True, tbx_logdir=f'{env_name}_log',\n",
    "  tbx_experiment_name=\"obs_hist_wrapped\", reward_scaling=1, \n",
    "  num_eval_envs = 64, num_envs=num_envs,\n",
    "  unroll_length=unroll_length, critic_batch_size=critic_batch_size,\n",
    "  critic_epochs = 16,target_critic_alpha = 0.2,\n",
    "  discounting = 0.99, lambda_ = 0.95,\n",
    "  normalize_observations=True,\n",
    "  episode_length=episode_length, \n",
    "  network_factory = make_networks_factory, \n",
    "  seed=10, entropy_cost = 1e-3,\n",
    "  resample_init=True,\n",
    "  scramble_initial_times=True,\n",
    "  num_grad_checks=num_grad_checks,\n",
    "  save_all_checkpoints=True,\n",
    "  save_all_policy_gradients=save_all_policy_gradients\n",
    ")\n",
    "\n",
    "trainer = SHAC(environment=env, eval_env=eval_env, **kwargs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial eval time: 8.7690 s\n",
      "Deleting old checkpoints!\n",
      "Checkpointed for epoch 0\n",
      "Checkpointed for epoch 1\n",
      "Checkpointed for epoch 2\n",
      "Checkpointed for epoch 3\n",
      "Checkpointed for epoch 4\n",
      "Checkpointed for epoch 5\n",
      "Checkpointed for epoch 6\n",
      "Checkpointed for epoch 7\n",
      "Checkpointed for epoch 8\n",
      "Checkpointed for epoch 9\n",
      "Checkpointed for epoch 10\n",
      "Checkpointed for epoch 11\n",
      "Checkpointed for epoch 12\n",
      "Checkpointed for epoch 13\n",
      "Checkpointed for epoch 14\n",
      "Checkpointed for epoch 15\n",
      "Checkpointed for epoch 16\n",
      "Checkpointed for epoch 17\n",
      "Checkpointed for epoch 18\n",
      "Checkpointed for epoch 19\n",
      "Checkpointed for epoch 20\n",
      "Checkpointed for epoch 21\n",
      "Checkpointed for epoch 22\n",
      "Checkpointed for epoch 23\n",
      "Checkpointed for epoch 24\n",
      "Checkpointed for epoch 25\n",
      "Checkpointed for epoch 26\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 27\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 28\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 29\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 31\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 32\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 33\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 34\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 35\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 37\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 38\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 39\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n",
      "WARNING:tensorboardX.x2num:NaN or Inf found in input tensor.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "at mapped index 0: Gradient ~ 0! (`check` failed)\n",
      "Checkpointed for epoch 40\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[9], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m make_inference_fn, policy_params, value_params, _\u001b[38;5;241m=\u001b[39m \u001b[43mtrainer\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtrain\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/Documents/thesis/jax_shac/shac/train_class.py:837\u001b[0m, in \u001b[0;36mSHAC.train\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    833\u001b[0m epoch_start_env_time \u001b[38;5;241m=\u001b[39m env_state\u001b[38;5;241m.\u001b[39mpipeline_state\u001b[38;5;241m.\u001b[39mtime[\u001b[38;5;241m0\u001b[39m]\n\u001b[1;32m    834\u001b[0m epoch_start_state \u001b[38;5;241m=\u001b[39m env_state\u001b[38;5;241m.\u001b[39mpipeline_state\u001b[38;5;241m.\u001b[39mqpos[\u001b[38;5;241m0\u001b[39m, \u001b[38;5;241m0\u001b[39m]\n\u001b[1;32m    836\u001b[0m (training_state, env_state,\n\u001b[0;32m--> 837\u001b[0m     training_metrics) \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtraining_epoch_with_timing\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtraining_state\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43menv_state\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mepoch_key\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    839\u001b[0m \u001b[38;5;66;03m# CHECKPOINT THE KEYS THAT WERE USED\u001b[39;00m\n\u001b[1;32m    840\u001b[0m algo_state[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124munroll_keys\u001b[39m\u001b[38;5;124m\"\u001b[39m] \u001b[38;5;241m=\u001b[39m training_metrics[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mtraining/unroll_keys\u001b[39m\u001b[38;5;124m'\u001b[39m][\u001b[38;5;241m0\u001b[39m] \u001b[38;5;66;03m# Remove extra batch dim.\u001b[39;00m\n",
      "File \u001b[0;32m~/Documents/thesis/jax_shac/shac/train_class.py:689\u001b[0m, in \u001b[0;36mSHAC.training_epoch_with_timing\u001b[0;34m(self, training_state, env_state, key)\u001b[0m\n\u001b[1;32m    685\u001b[0m \u001b[38;5;250m\u001b[39m\u001b[38;5;124;03m\"\"\" \u001b[39;00m\n\u001b[1;32m    686\u001b[0m \u001b[38;5;124;03mWrapper around training_epoch to time it.\u001b[39;00m\n\u001b[1;32m    687\u001b[0m \u001b[38;5;124;03m\"\"\"\u001b[39;00m\n\u001b[1;32m    688\u001b[0m t \u001b[38;5;241m=\u001b[39m time\u001b[38;5;241m.\u001b[39mtime()\n\u001b[0;32m--> 689\u001b[0m training_state, env_state, metrics, err \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtraining_epoch\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    690\u001b[0m \u001b[43m    \u001b[49m\u001b[43mtraining_state\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43menv_state\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mkey\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    692\u001b[0m jax\u001b[38;5;241m.\u001b[39mtree_util\u001b[38;5;241m.\u001b[39mtree_map(\u001b[38;5;28;01mlambda\u001b[39;00m x: x\u001b[38;5;241m.\u001b[39mblock_until_ready(), metrics)\n\u001b[1;32m    694\u001b[0m \u001b[38;5;66;03m#### CHECKIFYING\u001b[39;00m\n\u001b[1;32m    695\u001b[0m \u001b[38;5;66;03m# err.throw()\u001b[39;00m\n",
      "File \u001b[0;32m<string>:1\u001b[0m, in \u001b[0;36m<lambda>\u001b[0;34m(_cls)\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "make_inference_fn, policy_params, value_params, _= trainer.train()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Eval"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# utils_path = Path(Path.cwd().parent,\n",
    "#                   Path(\"benchmarks\"),\n",
    "#                   Path(\"quadruped_stability_studies\"))\n",
    "# sys.path.append(str(utils_path))\n",
    "\n",
    "from jax_shac.utils.eval_utils import plot_jacobian_norms, visualize_rollout, plot_rews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "renderer = mujoco.Renderer(env.model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "jit_reset = jax.jit(env.reset)\n",
    "jit_step = jax.jit(env.step)\n",
    "# jit_inference_fn = jax.jit(make_inference_fn(policy_params))\n",
    "\n",
    "rng = jax.random.key(141)\n",
    "init_key, rollout_key, rng = jax.random.split(rng, 3)\n",
    "s0 = jit_reset(init_key)\n",
    "\n",
    "jit_inference_fn = jax.jit(make_inference_fn(policy_params))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if False:\n",
    "    def spoof_policy(_, rng):\n",
    "        bb = jax.random.choice(rng, jp.array([-1, 1]), \n",
    "                            p=jp.array([0.9, 1]))\n",
    "        return bb.reshape(1), {}\n",
    "        # return 2*(jax.random.uniform(rng).reshape(1)-0.5), {}\n",
    "    jit_inference_fn = jax.jit(spoof_policy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x700 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if True:\n",
    "    \"\"\" \n",
    "    Nan gradients:\n",
    "    - Not seen in forward-mode AD\n",
    "    - Jacrev doesn't work\n",
    "    - Unclear whether nan's in training come from user code or MJX\n",
    "        - https://github.com/google-deepmind/mujoco/issues/1349\n",
    "    \"\"\"\n",
    "    all_jacs, state_jacs, ctrl_jacs = plot_jacobian_norms(s0, rollout_key, env, jit_inference_fn,\n",
    "                                                jit_step, episode_length=240,\n",
    "                                                ret_jacs=True, jacmode='fwd')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"show_videos\" style=\"border-spacing:0px;\"><tr><td style=\"padding:1px;\"><video controls width=\"320\" height=\"240\" style=\"object-fit:cover;\" loop autoplay muted>\n",
       "      <source src=\"data:video/mp4;base64,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\" type=\"video/mp4\"/>\n",
       "      This browser does not support the video tag.\n",
       "      </video></td></tr></table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "state_h, states = visualize_rollout(s0, rollout_key, env, \n",
    "                                    jit_inference_fn, jit_step, renderer, \n",
    "                                    episode_length=240, print_nans=False, \n",
    "                                    print_dones=True, ret_states=True, camera=\"track\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rews = plot_rews(states)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Debugging the Nans\n",
    "\n",
    "Below experiment run with: \n",
    "- action_pad = 0\n",
    "- target_critic_alpha = 0.2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Distribution of Nans in the policy gradient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Number of nan gradients')"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import jax.numpy as jp\n",
    "import pickle\n",
    "import jax\n",
    "import operator\n",
    "\n",
    "it = 24\n",
    "# Load policy gradients\n",
    "bgrad_path = f\"/home/andrew/Documents/thesis/jax_shac/shac/policy_gradients/bgrad_{it}.pkl\" # TODO\n",
    "bgrad = pickle.load(open(bgrad_path, \"rb\"))\n",
    "\n",
    "# Count the # nans across env axis\n",
    "app_nan = jax.tree_util.tree_map(lambda l: jp.isnan(l), bgrad)\n",
    "reduced_leaves = jax.tree_util.tree_map(lambda l: jp.sum(l, axis=(range(1, len(l.shape)))), app_nan)\n",
    "nans_per_env = jax.tree.reduce(operator.add, reduced_leaves)\n",
    "# nan_mask = jnp.isnan(bgrad)\n",
    "# nans_per_env = jnp.sum(nan_mask, axis=0)\n",
    "# # Plot\n",
    "num_envs = 64\n",
    "\n",
    "plt.scatter(jp.linspace(0, num_envs-1, num_envs), nans_per_env)\n",
    "plt.xlabel(\"Environment Number\")\n",
    "plt.ylabel(\"Number of nan gradients\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Correspondence between an environment having nan policy gradient, and it having at least one nan mjx jacobian\n",
    "\n",
    "Conclusion:\n",
    "- Having at least one nan in a state jacobian is sufficient to cause a nan policy gradient\n",
    "- There are other causes of nan policy gradients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(64, 32, 6, 7)\n"
     ]
    }
   ],
   "source": [
    "# Collect a jacobian rollout (13 sec first run)\n",
    "\n",
    "algo_state = trainer.load_checkpoint(it)\n",
    "ts = algo_state['training_state']\n",
    "state = algo_state['env_state']\n",
    "key = algo_state['unroll_keys']\n",
    "\n",
    "jacs, nstates, extras = trainer.jac_rollout(ts.policy_params, ts.normalizer_params,\n",
    "                           state, key)\n",
    "actions = extras[\"action\"]\n",
    "print(jacs.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fa1f97d62f0>"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pg_nan_mask = nans_per_env > 0\n",
    "\n",
    "# See which envs had at least one nan.\n",
    "jacs_nan_mask = jp.sum(jp.isnan(jacs), axis=(1,2,3))\n",
    "jacs_nan_mask = jacs_nan_mask > 0\n",
    "\n",
    "x_axis = jp.linspace(0, 63, 64)\n",
    "\n",
    "plt.plot(pg_nan_mask, label=\"Nan policy gradient\")\n",
    "plt.plot(jacs_nan_mask, label=\"Nan jacobian\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0.98, 'Number of Nan Indicies; Brighter is Larger')"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\" \n",
    "Take-aways: \n",
    "- Either the entire jacobian is nan or none of it.\n",
    "- Nans are randomly located along the trajectory.\n",
    "\"\"\"\n",
    "\n",
    "# Where are the nan's happening?\n",
    "state_dim = env.sys.nq + env.sys.nv\n",
    "action_dim = env.sys.nu\n",
    "\n",
    "state_jacs = jacs[..., :state_dim]\n",
    "action_jacs  = jacs[..., state_dim:state_dim+action_dim]\n",
    "\n",
    "sjacs_nan_mask = jp.sum(jp.isnan(jacs), axis=(2,3))\n",
    "ajacs_nan_mask = jp.sum(jp.isnan(jacs), axis=(2,3))\n",
    "\n",
    "fig, axs = plt.subplots(nrows=1, ncols=2, figsize=(10, 5))\n",
    "axs[0].imshow(sjacs_nan_mask, aspect='auto')\n",
    "axs[1].imshow(ajacs_nan_mask, aspect='auto')\n",
    "axs[0].set_title(\"State jacobian\")\n",
    "axs[1].set_title(\"Action jacobian\")\n",
    "axs[0].set_xlabel(\"Env step\")\n",
    "axs[0].set_ylabel(\"Env number\")\n",
    "axs[1].set_xlabel(\"Env step\")\n",
    "axs[1].set_ylabel(\"Env number\")\n",
    "# fig.text(0.5, 1, \"hi\", )\n",
    "fig.suptitle(\"Number of Nan Indicies; Brighter is Larger\", size='x-large')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'What are the states and actions resulting in nan jacobians?')"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# WLOG\n",
    "n_actions = actions[sjacs_nan_mask>1]\n",
    "n_qpos = nstates.pipeline_state.qpos[sjacs_nan_mask>1]\n",
    "\n",
    "plt.scatter(n_actions, n_qpos[:, 1], label='Height')\n",
    "plt.scatter(n_actions, n_qpos[:, 2], label=\"Pole position\")\n",
    "plt.legend()\n",
    "plt.xlabel(\"Action\")\n",
    "plt.title(\"What are the states and actions resulting in nan jacobians?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Array([-147.99928], dtype=float32)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(n_actions)\n",
    "# n_actions.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Distribution of actions (envs, time)')"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "flat_actions = jp.reshape(actions, newshape=(actions.size,))\n",
    "plt.hist(flat_actions)\n",
    "plt.title(\"Distribution of actions (envs, time)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([4., 4., 5., 2., 5., 7., 2., 2., 5., 1.]),\n",
       " array([0.54000014, 0.95200282, 1.36400557, 1.77600825, 2.18801093,\n",
       "        2.60001349, 3.0120163 , 3.42401886, 3.83602166, 4.24802446,\n",
       "        4.66002703]),\n",
       " <BarContainer object of 10 artists>)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Does it have to do with time?\n",
    "plt.hist(nstates.pipeline_state.time[sjacs_nan_mask>1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Array([[3.0800319 , 3.1000328 , 3.1200337 , ..., 3.660059  , 3.68006   ,\n",
       "        3.7000608 ],\n",
       "       [1.6799855 , 1.6999853 , 1.719985  , ..., 2.2599936 , 2.2799945 ,\n",
       "        2.2999954 ],\n",
       "       [3.5200524 , 3.5400534 , 3.5600543 , ..., 4.1000676 , 4.120066  ,\n",
       "        4.1400647 ],\n",
       "       ...,\n",
       "       [1.9799817 , 1.9999814 , 2.0199823 , ..., 2.5600076 , 2.5800085 ,\n",
       "        2.6000094 ],\n",
       "       [1.879983  , 1.8999827 , 1.9199824 , ..., 2.460003  , 2.4800038 ,\n",
       "        2.5000048 ],\n",
       "       [0.18      , 0.20000005, 0.22000009, ..., 0.7599973 , 0.77999705,\n",
       "        0.7999968 ]], dtype=float32, weak_type=True)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nstates.pipeline_state.time"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "mujoco",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}