Part 4 build_makemore_mlp2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "2f4f6809-b585-472f-8629-988b86aca95f",
   "metadata": {},
   "source": [
    "# makemore: part 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cd42fb7e-f785-4762-84ed-eaecde6833a9",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn.functional as F\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "77f98d0d-1058-43b4-bb50-1da72d06d8c9",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['emma', 'olivia', 'ava', 'isabella', 'sophia', 'charlotte', 'mia', 'amelia']"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "words = open('names.txt', 'r').read().splitlines()\n",
    "words[:8]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "9310d494-aeb8-4c93-8e98-4313cb3120c4",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{1: 'a', 2: 'b', 3: 'c', 4: 'd', 5: 'e', 6: 'f', 7: 'g', 8: 'h', 9: 'i', 10: 'j', 11: 'k', 12: 'l', 13: 'm', 14: 'n', 15: 'o', 16: 'p', 17: 'q', 18: 'r', 19: 's', 20: 't', 21: 'u', 22: 'v', 23: 'w', 24: 'x', 25: 'y', 26: 'z', 0: '.'}\n",
      "27\n"
     ]
    }
   ],
   "source": [
    "chars = sorted(list(set(''.join(words))))\n",
    "stoi = {s:i+1 for i,s in enumerate(chars)}\n",
    "stoi['.'] = 0\n",
    "itos = {i:s for s,i in stoi.items()}\n",
    "vocab_size = len(itos)\n",
    "print(itos)\n",
    "print(vocab_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "312d7370-f99c-4e61-a08f-e2052cd293dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([182625, 3]) torch.Size([182625])\n",
      "torch.Size([22655, 3]) torch.Size([22655])\n",
      "torch.Size([22866, 3]) torch.Size([22866])\n"
     ]
    }
   ],
   "source": [
    "# build dataset\n",
    "block_size = 3\n",
    "\n",
    "def build_dataset(words):\n",
    "    X, Y = [], []\n",
    "    \n",
    "    for w in words:\n",
    "        context = [0] * block_size\n",
    "        for ch in w + '.':\n",
    "            ix = stoi[ch]\n",
    "            X.append(context)\n",
    "            Y.append(ix)\n",
    "            context = context[1:] + [ix]\n",
    "            \n",
    "    X = torch.tensor(X)\n",
    "    Y = torch.tensor(Y)\n",
    "    print(X.shape, Y.shape)\n",
    "    return X, Y\n",
    "\n",
    "import random\n",
    "random.seed(42)\n",
    "random.shuffle(words)\n",
    "n1 = int(0.8*len(words))\n",
    "n2 = int(0.9*len(words))\n",
    "\n",
    "Xtr, Ytr = build_dataset(words[:n1])\n",
    "Xdev, Ydev = build_dataset(words[n1:n2])\n",
    "Xte, Yte = build_dataset(words[n2:])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5cdf898f-4395-494a-b670-83ad9ca5b4e9",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12297\n"
     ]
    }
   ],
   "source": [
    "# MLP revisited\n",
    "n_embd = 10\n",
    "n_hidden = 200\n",
    "\n",
    "g = torch.Generator().manual_seed(2147483647)\n",
    "C = torch.randn((vocab_size, n_embd), generator=g)\n",
    "W1 = torch.randn((n_embd * block_size, n_hidden), generator=g) * 0.2\n",
    "b1 = torch.randn(n_hidden, generator=g) * 0.01\n",
    "W2 = torch.randn((n_hidden, vocab_size), generator=g) * 0.01\n",
    "b2 = torch.randn(vocab_size, generator=g) * 0.0\n",
    "\n",
    "bngain = torch.ones((1, n_hidden))\n",
    "bnbias = torch.zeros((1, n_hidden))\n",
    "\n",
    "parameters = [C, W1, b1, W2, b2, bngain, bnbias]\n",
    "print(sum(p.nelement() for p in parameters))\n",
    "for p in parameters:\n",
    "    p.requires_grad = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "fca38a15-bde7-4644-baec-d5bfc5eae7c2",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      0/ 200000: 3.3147\n",
      "  10000/ 200000: 2.1341\n",
      "  20000/ 200000: 2.3173\n",
      "  30000/ 200000: 2.4115\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[7], line 20\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m p \u001b[38;5;129;01min\u001b[39;00m parameters:\n\u001b[1;32m     19\u001b[0m     p\u001b[38;5;241m.\u001b[39mgrad \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[0;32m---> 20\u001b[0m loss\u001b[38;5;241m.\u001b[39mbackward()\n\u001b[1;32m     22\u001b[0m lr \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m0.1\u001b[39m \u001b[38;5;28;01mif\u001b[39;00m i \u001b[38;5;241m<\u001b[39m \u001b[38;5;241m100000\u001b[39m \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;241m0.01\u001b[39m\n\u001b[1;32m     23\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m p \u001b[38;5;129;01min\u001b[39;00m parameters:\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/torch/_tensor.py:487\u001b[0m, in \u001b[0;36mTensor.backward\u001b[0;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[1;32m    477\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m has_torch_function_unary(\u001b[38;5;28mself\u001b[39m):\n\u001b[1;32m    478\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m handle_torch_function(\n\u001b[1;32m    479\u001b[0m         Tensor\u001b[38;5;241m.\u001b[39mbackward,\n\u001b[1;32m    480\u001b[0m         (\u001b[38;5;28mself\u001b[39m,),\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    485\u001b[0m         inputs\u001b[38;5;241m=\u001b[39minputs,\n\u001b[1;32m    486\u001b[0m     )\n\u001b[0;32m--> 487\u001b[0m torch\u001b[38;5;241m.\u001b[39mautograd\u001b[38;5;241m.\u001b[39mbackward(\n\u001b[1;32m    488\u001b[0m     \u001b[38;5;28mself\u001b[39m, gradient, retain_graph, create_graph, inputs\u001b[38;5;241m=\u001b[39minputs\n\u001b[1;32m    489\u001b[0m )\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/torch/autograd/__init__.py:200\u001b[0m, in \u001b[0;36mbackward\u001b[0;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[1;32m    195\u001b[0m     retain_graph \u001b[38;5;241m=\u001b[39m create_graph\n\u001b[1;32m    197\u001b[0m \u001b[38;5;66;03m# The reason we repeat same the comment below is that\u001b[39;00m\n\u001b[1;32m    198\u001b[0m \u001b[38;5;66;03m# some Python versions print out the first line of a multi-line function\u001b[39;00m\n\u001b[1;32m    199\u001b[0m \u001b[38;5;66;03m# calls in the traceback and some print out the last line\u001b[39;00m\n\u001b[0;32m--> 200\u001b[0m Variable\u001b[38;5;241m.\u001b[39m_execution_engine\u001b[38;5;241m.\u001b[39mrun_backward(  \u001b[38;5;66;03m# Calls into the C++ engine to run the backward pass\u001b[39;00m\n\u001b[1;32m    201\u001b[0m     tensors, grad_tensors_, retain_graph, create_graph, inputs,\n\u001b[1;32m    202\u001b[0m     allow_unreachable\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m, accumulate_grad\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "max_steps = 200000\n",
    "batch_size = 32\n",
    "lossi = []\n",
    "\n",
    "for i in range(max_steps):\n",
    "    \n",
    "    ix = torch.randint(0, Xtr.shape[0], (batch_size, ), generator=g)\n",
    "    Xb, Yb = Xtr[ix], Ytr[ix]\n",
    "    \n",
    "    emb = C[Xb]\n",
    "    embcat = emb.view(emb.shape[0], -1)\n",
    "    hpreact = embcat @ W1 + b1\n",
    "    hpreact = bngain * (hpreact - hpreact.mean(0, keepdim=True)) / hpreact.std(0, keepdim=True) + bnbias\n",
    "    h = torch.tanh(hpreact)\n",
    "    logits = h @ W2 + b2\n",
    "    loss = F.cross_entropy(logits, Yb)\n",
    "    \n",
    "    for p in parameters:\n",
    "        p.grad = None\n",
    "    loss.backward()\n",
    "    \n",
    "    lr = 0.1 if i < 100000 else 0.01\n",
    "    for p in parameters:\n",
    "        p.data += -lr * p.grad\n",
    "    \n",
    "    if i % 10000 == 0: \n",
    "        print(f'{i:7d}/{max_steps:7d}: {loss.item():.4f}')\n",
    "    lossi.append(loss.log10().item())\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2c80eebf-62cb-42a1-9cd4-35b6575cf3f4",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x15f377b90>]"
      ]
     },
     "execution_count": 7,
     "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": [
    "plt.plot(lossi)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2f5f398d-e5d5-4567-a64c-7ee5cff68e71",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'bnmean' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[8], line 18\u001b[0m\n\u001b[1;32m     15\u001b[0m     loss \u001b[38;5;241m=\u001b[39m F\u001b[38;5;241m.\u001b[39mcross_entropy(logits, y)\n\u001b[1;32m     16\u001b[0m     \u001b[38;5;28mprint\u001b[39m(split, loss\u001b[38;5;241m.\u001b[39mitem())\n\u001b[0;32m---> 18\u001b[0m split_loss(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mtrain\u001b[39m\u001b[38;5;124m'\u001b[39m)\n\u001b[1;32m     19\u001b[0m split_loss(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mval\u001b[39m\u001b[38;5;124m'\u001b[39m)\n",
      "File \u001b[0;32m~/anaconda3/lib/python3.11/site-packages/torch/utils/_contextlib.py:115\u001b[0m, in \u001b[0;36mcontext_decorator.<locals>.decorate_context\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    112\u001b[0m \u001b[38;5;129m@functools\u001b[39m\u001b[38;5;241m.\u001b[39mwraps(func)\n\u001b[1;32m    113\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mdecorate_context\u001b[39m(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs):\n\u001b[1;32m    114\u001b[0m     \u001b[38;5;28;01mwith\u001b[39;00m ctx_factory():\n\u001b[0;32m--> 115\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m func(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)\n",
      "Cell \u001b[0;32mIn[8], line 12\u001b[0m, in \u001b[0;36msplit_loss\u001b[0;34m(split)\u001b[0m\n\u001b[1;32m     10\u001b[0m hpreact \u001b[38;5;241m=\u001b[39m embcat \u001b[38;5;241m@\u001b[39m W1 \u001b[38;5;241m+\u001b[39m b1\n\u001b[1;32m     11\u001b[0m \u001b[38;5;66;03m#hpreact = bngain * (hpreact - hpreact.mean(0, keepdim=True)) / hpreact.std(0, keepdim=True) + bnbias\u001b[39;00m\n\u001b[0;32m---> 12\u001b[0m hpreact \u001b[38;5;241m=\u001b[39m bngain \u001b[38;5;241m*\u001b[39m (hpreact \u001b[38;5;241m-\u001b[39m bnmean) \u001b[38;5;241m/\u001b[39m bnstd \u001b[38;5;241m+\u001b[39m bnbias\n\u001b[1;32m     13\u001b[0m h \u001b[38;5;241m=\u001b[39m torch\u001b[38;5;241m.\u001b[39mtanh(hpreact)\n\u001b[1;32m     14\u001b[0m logits \u001b[38;5;241m=\u001b[39m h \u001b[38;5;241m@\u001b[39m W2 \u001b[38;5;241m+\u001b[39m b2\n",
      "\u001b[0;31mNameError\u001b[0m: name 'bnmean' is not defined"
     ]
    }
   ],
   "source": [
    "@torch.no_grad()\n",
    "def split_loss(split):\n",
    "    x, y = {\n",
    "        'train': (Xtr, Ytr),\n",
    "        'val': (Xdev, Ydev),\n",
    "        'test': (Xte, Yte)\n",
    "    }[split]\n",
    "    emb = C[x]\n",
    "    embcat = emb.view(emb.shape[0], -1)\n",
    "    hpreact = embcat @ W1 + b1\n",
    "    #hpreact = bngain * (hpreact - hpreact.mean(0, keepdim=True)) / hpreact.std(0, keepdim=True) + bnbias\n",
    "    hpreact = bngain * (hpreact - bnmean) / bnstd + bnbias\n",
    "    h = torch.tanh(hpreact)\n",
    "    logits = h @ W2 + b2\n",
    "    loss = F.cross_entropy(logits, y)\n",
    "    print(split, loss.item())\n",
    "\n",
    "split_loss('train')\n",
    "split_loss('val')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca10ceab-1c49-4dc9-ae7c-7d338b204ed0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "70329b06-f84a-4f71-a54e-707ae11b9c81",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([0.1749, 0.4754, 0.1749, 0.1749]), tensor(1.7437))"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 4d example of the issue (mis-initialization at start)\n",
    "logits = torch.tensor([0.0, 1.0, 0, 0.0])\n",
    "probs = torch.softmax(logits, dim=0)\n",
    "loss = -probs[2].log()\n",
    "probs, loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "1ab17664-ab47-45f0-a584-106f4a5177c5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(h.view(-1).tolist(), 50);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "423b790f-9aee-4d27-8333-a92b1862b70a",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(0.0047) tensor(0.9920)\n",
      "tensor(0.0082) tensor(3.1722)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = torch.randn(1000, 10)\n",
    "w = torch.randn(10, 200)\n",
    "y = x @ w\n",
    "print(x.mean(), x.std())\n",
    "print(y.mean(), y.std())\n",
    "plt.figure(figsize=(20,5))\n",
    "plt.subplot(121)\n",
    "plt.hist(x.view(-1).tolist(), 50, density=True);\n",
    "plt.subplot(122)\n",
    "plt.hist(y.view(-1).tolist(), 50, density=True);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2dee22b7-8c43-424b-8abc-b50878267839",
   "metadata": {},
   "source": [
    "batchnormalization\n",
    "makes samples (neuros and their logits) dependent on each other, in a random fashion\n",
    "forward pass: network expects batches now (not single examples)\n",
    "resolution: calibrate batchnormalization on the entire training set (get fixed pvalues)\n",
    "initialization want everything roughly Gaussian\n",
    "tanh don't want input values too large, otherwise gradient is essentially 0\n",
    "want to check that initialization gradients are not too close to 0, otherwise nothing will be learned\n",
    "running mean and std of batchnorm, instead of one entire forward pass to get the training mean and std\n",
    "see also pytorch linear layer, batchnormalization layer\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0fb169a3-348f-48c8-812d-941cce324716",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Calibrate the batch norm at the end of training\n",
    "\n",
    "with torch.no_grad():\n",
    "    # pass the training set through\n",
    "    emb = C[Xtr]\n",
    "    embcat = emb.view(emb.shape[0], -1)\n",
    "    hpreact = embcat @ W1 + b1\n",
    "    bnmean = hpreact.mean(0, keepdim=True)\n",
    "    bnstd = hpreact.std(0, keepdim=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b32dc37b-5356-489d-9ae4-ccd41bfda5b9",
   "metadata": {},
   "source": [
    "# Pytorchifying"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "24b0ffcb-fe88-43c6-8835-67a93758b90c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Let's train a deeper network\n",
    "\n",
    "class Linear:\n",
    "    \n",
    "    def __init__(self, fan_in, fan_out, bias=True):\n",
    "        # self.weight = torch.randn((fan_in, fan_out), generator=g) / fan_in**0.5\n",
    "        self.weight = torch.zeros((fan_in, fan_out)) / fan_in**0.5\n",
    "        self.bias = torch.zeros(fan_out) if bias else None\n",
    "    \n",
    "    def __call__(self, x):\n",
    "        self.out = x @ self.weight\n",
    "        if self.bias is not None:\n",
    "            self.out += self.bias\n",
    "        return self.out\n",
    "    \n",
    "    def parameters(self):\n",
    "        return [self.weight] + ([] if self.bias is None else [self.bias])\n",
    "\n",
    "class BatchNorm1d:\n",
    "    \n",
    "    def __init__(self, dim, eps=1e-5, momentum=0.1):\n",
    "        self.eps = eps\n",
    "        self.momentum = momentum\n",
    "        self.training = True\n",
    "        # parameters with backprop\n",
    "        self.gamma = torch.ones(dim)\n",
    "        self.beta = torch.zeros(dim)\n",
    "        # buffers (running update)\n",
    "        self.running_mean = torch.zeros(dim)\n",
    "        self.running_var = torch.ones(dim)\n",
    "    \n",
    "    def __call__(self, x):\n",
    "        # forward pass\n",
    "        if self.training:\n",
    "            xmean = x.mean(0, keepdim=True) # batch mean\n",
    "            xvar = x.var(0, keepdim=True) # batch variance\n",
    "        else:\n",
    "            xmean = self.running_mean\n",
    "            xvar = self.running_var\n",
    "        xhat = (x - xmean) / torch.sqrt(xvar + self.eps)\n",
    "        self.out = self.gamma * xhat + self.beta\n",
    "        # update buffers\n",
    "        if self.training:\n",
    "            with torch.no_grad():\n",
    "                self.running_mean = (1 - self.momentum) * self.running_mean + self.momentum * xmean\n",
    "                self.running_var = (1 - self.momentum) * self.running_var + self.momentum * xvar\n",
    "        return self.out\n",
    "    \n",
    "    def parameters(self):\n",
    "        return [self.gamma, self.beta]\n",
    "    \n",
    "class Tanh:\n",
    "    def __call__(self, x):\n",
    "        self.out = torch.tanh(x)\n",
    "        return self.out\n",
    "    def parameters(self):\n",
    "        return []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "4d64fb4d-9d3d-44be-a080-4b117f77ea6e",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "46497\n"
     ]
    }
   ],
   "source": [
    "n_embd = 10\n",
    "n_hidden = 100\n",
    "g = torch.Generator().manual_seed(2147483647)\n",
    "\n",
    "C = torch.randn((vocab_size, n_embd), generator=g)\n",
    "layers = [\n",
    "    Linear(n_embd * block_size, n_hidden), Tanh(),\n",
    "    Linear(n_hidden,            n_hidden), Tanh(),\n",
    "    Linear(n_hidden,            n_hidden), Tanh(),\n",
    "    Linear(n_hidden,            n_hidden), Tanh(),\n",
    "    Linear(n_hidden,            n_hidden), Tanh(),\n",
    "    Linear(n_hidden, vocab_size), \n",
    "]\n",
    "\n",
    "with torch.no_grad():\n",
    "    layers[-1].weight += 0.1 # \"less confident\" not for batchnorm as last layer\n",
    "    # layers[-1].gamma += 0.1 # if batchnorm is last layer\n",
    "    # apply gain\n",
    "    for layer in layers[:-1]:\n",
    "        if isinstance(layer, Linear):\n",
    "            # may need to update learning rate\n",
    "            layer.weight *= 5/3 # \"gain\" so that output std of tanh \"squashing\" remains stable under many iterations\n",
    "\n",
    "parameters = [C] + [p for layer in layers for p in layer.parameters()]\n",
    "\n",
    "print(sum(p.nelement() for p in parameters))\n",
    "\n",
    "for p in parameters:\n",
    "    p.requires_grad = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "b1f568b2-8036-4e01-bc6f-6cad17d2e410",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      0/ 200000: 2.7505\n",
      "  10000/ 200000: 2.9572\n",
      "  20000/ 200000: 3.1180\n",
      "  30000/ 200000: 2.8065\n"
     ]
    }
   ],
   "source": [
    "# same as before\n",
    "max_steps = 200000\n",
    "batch_size = 32\n",
    "lossi = []\n",
    "ud = [] # update to data ratio\n",
    "for i in range(max_steps):\n",
    "    \n",
    "    # minibatch\n",
    "    ix = torch.randint(0, Xtr.shape[0], (batch_size, ), generator=g)\n",
    "    Xb, Yb = Xtr[ix], Ytr[ix]\n",
    "    \n",
    "    # forward pass\n",
    "    emb = C[Xb]\n",
    "    x = emb.view(emb.shape[0], -1)\n",
    "    for layer in layers:\n",
    "        x  = layer(x)\n",
    "    loss = F.cross_entropy(x, Yb)\n",
    "    \n",
    "    # backward pass\n",
    "    for layer in layers:\n",
    "        layer.out.retain_grad() # AFTER_DEBUG: retain_graph() ? own layers don't have out, or this function?\n",
    "    for p in parameters:\n",
    "        p.grad = None\n",
    "    loss.backward()\n",
    "    \n",
    "    # update\n",
    "    lr = 0.1 if i < 100000 else 0.01 # check this is the update ratio plot\n",
    "    for p in parameters:\n",
    "        p.data += -lr * p.grad\n",
    "    \n",
    "    # stats\n",
    "    if i % 10000 == 0:\n",
    "        print(f'{i:7d}/{max_steps:7d}: {loss.item():.4f}')\n",
    "    lossi.append(loss.log10().item())\n",
    "    with torch.no_grad():\n",
    "        ud.append([(lr*p.grad.std() / p.data.std()).log10().item() for p in parameters])\n",
    "    \n",
    "    if i >= 30001:\n",
    "        break # AFTER_DEBUG take out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "9f7daef0-8a23-4188-bbc1-b6b9b1a6bc86",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0., 1., 2., 3., 4., 5., 6., 7., 8., 9.],\n",
       "        [0., 1., 2., 3., 4., 5., 6., 7., 8., 9.]], dtype=torch.float64)"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.tensor([range(10), range(10)], dtype=float).data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "d2bb43b9-1c57-4294-814a-c7a366eac928",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "layer 1 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 3 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 5 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 7 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 9 (      Tanh): mean -0.00, std 0.00, saturated: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize histograms\n",
    "# layer output distribution\n",
    "plt.figure(figsize=(20, 4))\n",
    "legends = []\n",
    "for i, layer in enumerate(layers[:-1]): # not over the output layer\n",
    "    if isinstance(layer, Tanh):\n",
    "        t = layer.out\n",
    "        print('layer %d (%10s): mean %+.2f, std %.2f, saturated: %.2f%%' \n",
    "              % (i, layer.__class__.__name__, t.mean(), t.std(), \n",
    "                 (t.abs() > 0.97).float().mean()*100))\n",
    "        hy, hx = torch.histogram(t, density=True)\n",
    "        plt.plot(hx[:-1].detach(), hy.detach())\n",
    "        legends.append(f'layer {i} ({layer.__class__.__name__})')\n",
    "    plt.legend(legends);\n",
    "    plt.title('activation distribution')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "cbf3520e-4dcc-4881-98c1-47d95ff5d313",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "layer 1 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 3 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 5 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 7 (      Tanh): mean +0.00, std 0.00, saturated: 0.00%\n",
      "layer 9 (      Tanh): mean -0.00, std 0.00, saturated: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize histograms\n",
    "# layer grad distribution\n",
    "plt.figure(figsize=(20, 4))\n",
    "legends = []\n",
    "for i, layer in enumerate(layers[:-1]): # not over the output layer\n",
    "    if isinstance(layer, Tanh):\n",
    "        t = layer.out.grad\n",
    "        print('layer %d (%10s): mean %+.2f, std %.2f, saturated: %.2f%%' \n",
    "              % (i, layer.__class__.__name__, t.mean(), t.std(), \n",
    "                 (t.abs() > 0.97).float().mean()*100))\n",
    "        hy, hx = torch.histogram(t, density=True)\n",
    "        plt.plot(hx[:-1].detach(), hy.detach())\n",
    "        legends.append(f'layer {i} ({layer.__class__.__name__})')\n",
    "    plt.legend(legends);\n",
    "    plt.title('activation distribution')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "3341de70-d094-438a-8998-be273c9e9dba",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "weight   (27, 10) | mean +0.000000 | std 0.000000e+00 | grad:data ratio 0.000000e+00\n",
      "weight  (30, 100) | mean +0.000000 | std 0.000000e+00 | grad:data ratio nan\n",
      "weight (100, 100) | mean +0.000000 | std 0.000000e+00 | grad:data ratio nan\n",
      "weight (100, 100) | mean +0.000000 | std 0.000000e+00 | grad:data ratio nan\n",
      "weight (100, 100) | mean +0.000000 | std 0.000000e+00 | grad:data ratio nan\n",
      "weight (100, 100) | mean +0.000000 | std 0.000000e+00 | grad:data ratio nan\n",
      "weight  (100, 27) | mean +0.000000 | std 1.777247e-09 | grad:data ratio 1.806238e-03\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'activation distribution')"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize histograms\n",
    "# weights grad distribution\n",
    "plt.figure(figsize=(20, 4))\n",
    "legends = []\n",
    "for i, p in enumerate(parameters):\n",
    "    t = p.grad\n",
    "    if p.ndim == 2: # only matrices, no gammas, betas, biases\n",
    "        print('weight %10s | mean %+f | std %e | grad:data ratio %e' \n",
    "              % (tuple(p.shape), t.mean(), t.std(), t.std() / p.std()) )\n",
    "        hy, hx = torch.histogram(t, density=True)\n",
    "        plt.plot(hx[:-1].detach(), hy.detach())\n",
    "        legends.append(f'{i} {tuple(p.shape)}')\n",
    "plt.legend(legends)\n",
    "plt.title('activation distribution')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd66bec1-49fb-48ea-b7a5-0d10068ac60e",
   "metadata": {},
   "source": [
    "weights.grad.std() / weights.std() is a measure for the order of magnitude difference in the gradients vs the weights themselves (assuming they are centered around 0?), because later the weights are updated by learning rate x gradient. The gradients and existing weights should be of comparable relative order of magnitude across the layers (unless a layer-specific learning rate is used?). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e11ff5c-2dad-47fe-bd19-3b6c476e7d48",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "a99a6979-868e-4274-a4ef-4de48612a665",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x28457f450>]"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20, 4))\n",
    "t = torch.tanh(torch.randn(32*100)/ 32**0.5 * 5/3)\n",
    "hy, hx = torch.histogram(t, density=True)\n",
    "plt.plot(hx[:-1].detach(), hy.detach())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b01794fe-967c-4702-a8a1-f63e5fba6967",
   "metadata": {},
   "source": [
    "\"The backward gradient pass for a composition of linear layers is not the same as the backward pass for a single (equivalent) linear layer.\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "b3d70ba4-85f1-4906-add0-1f0b0ce829c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# one more plot on data update ratio\n",
    "plt.figure(figsize=(20, 4))\n",
    "legends = []\n",
    "for i,p in enumerate(parameters):\n",
    "    if p.ndim == 2:\n",
    "        plt.plot([ud[j][i] for j in range(len(ud))])\n",
    "        legends.append('param %d' % i)\n",
    "plt.plot([0, len(ud)], [-3, -3], 'k')\n",
    "plt.legend(legends);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "4d966201-18a7-4de0-bac2-f25427aaa0e3",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'Tanh' object has no attribute 'grad'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[113], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m layers[\u001b[38;5;241m-\u001b[39m\u001b[38;5;241m2\u001b[39m]\u001b[38;5;241m.\u001b[39mgrad\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'Tanh' object has no attribute 'grad'"
     ]
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "73ffd98d-d860-467e-8f61-a1cc40250d88",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}