build_makemore_mlp.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/julurisaichandu/makemore_andrej_karpathy/blob/master/build_makemore_mlp.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vowXRstocd8F"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import torch.nn.functional as F\n",
        "import matplotlib.pyplot as plt # for making figures\n",
        "%matplotlib inline"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# download the names.txt file from github\n",
        "!wget https://raw.githubusercontent.com/karpathy/makemore/master/names.txt"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9Rlc3YeWcm_d",
        "outputId": "cb8e6290-6428-4a2d-9e78-995c6e90f5c3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-09-11 02:48:38--  https://raw.githubusercontent.com/karpathy/makemore/master/names.txt\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.111.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 228145 (223K) [text/plain]\n",
            "Saving to: ‘names.txt’\n",
            "\n",
            "names.txt           100%[===================>] 222.80K  --.-KB/s    in 0.02s   \n",
            "\n",
            "2022-09-11 02:48:38 (11.4 MB/s) - ‘names.txt’ saved [228145/228145]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "words = open('names.txt', 'r').read().splitlines()\n",
        "words[:8]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uH2MGP4FcidR",
        "outputId": "7a128b70-343a-493a-85f4-3e80293695be"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['emma', 'olivia', 'ava', 'isabella', 'sophia', 'charlotte', 'mia', 'amelia']"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(words)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FbakoEqAcvnF",
        "outputId": "28507b40-8bab-4a62-d9c3-1152550ef0d4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "32033"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# build the vocabulary of characters and mappings to/from integers\n",
        "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",
        "print(itos)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EYnXhmWVcx-l",
        "outputId": "6dcad8cf-9caf-401d-e772-c26ac8f93f39"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "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"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# build the dataset\n",
        "block_size = 3 # context length: how many characters do we take to predict the next one?\n",
        "\n",
        "def build_dataset(words):\n",
        "  X, Y = [], []\n",
        "  for w in words:\n",
        "\n",
        "    #print(w)\n",
        "    context = [0] * block_size\n",
        "    for ch in w + '.':\n",
        "      ix = stoi[ch]\n",
        "      X.append(context)\n",
        "      Y.append(ix)\n",
        "      #print(''.join(itos[i] for i in context), '--->', itos[ix])\n",
        "      context = context[1:] + [ix] # crop and append\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:])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MB5XgJxdczFi",
        "outputId": "ce4e48c3-b9f7-4d1d-9e9a-47c37beb326e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([182580, 3]) torch.Size([182580])\n",
            "torch.Size([22767, 3]) torch.Size([22767])\n",
            "torch.Size([22799, 3]) torch.Size([22799])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "g = torch.Generator().manual_seed(2147483647) # for reproducibility\n",
        "C = torch.randn((27, 10), generator=g)\n",
        "W1 = torch.randn((30, 200), generator=g)\n",
        "b1 = torch.randn(200, generator=g)\n",
        "W2 = torch.randn((200, 27), generator=g)\n",
        "b2 = torch.randn(27, generator=g)\n",
        "parameters = [C, W1, b1, W2, b2]"
      ],
      "metadata": {
        "id": "EAA0_oigc13X"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sum(p.nelement() for p in parameters) # number of parameters in total"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_HcVef6-c5S4",
        "outputId": "fbcc9087-e5c6-4916-ba32-45812804bea0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "11897"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for p in parameters:\n",
        "  p.requires_grad = True"
      ],
      "metadata": {
        "id": "coNBea1sc6tm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lre = torch.linspace(-3, 0, 1000)\n",
        "lrs = 10**lre"
      ],
      "metadata": {
        "id": "nFWpAj_9c9Yr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lri = []\n",
        "lossi = []\n",
        "stepi = []"
      ],
      "metadata": {
        "id": "t2huuyTudBlF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(200000):\n",
        "\n",
        "  # minibatch construct\n",
        "  ix = torch.randint(0, Xtr.shape[0], (32,))\n",
        "\n",
        "  # forward pass\n",
        "  emb = C[Xtr[ix]] # (32, 3, 2)\n",
        "  h = torch.tanh(emb.view(-1, 30) @ W1 + b1) # (32, 100)\n",
        "  logits = h @ W2 + b2 # (32, 27)\n",
        "  loss = F.cross_entropy(logits, Ytr[ix])\n",
        "  #print(loss.item())\n",
        "\n",
        "  # backward pass\n",
        "  for p in parameters:\n",
        "    p.grad = None\n",
        "  loss.backward()\n",
        "\n",
        "  # update\n",
        "  #lr = lrs[i]\n",
        "  lr = 0.1 if i < 100000 else 0.01\n",
        "  for p in parameters:\n",
        "    p.data += -lr * p.grad\n",
        "\n",
        "  # track stats\n",
        "  #lri.append(lre[i])\n",
        "  stepi.append(i)\n",
        "  lossi.append(loss.log10().item())\n",
        "\n",
        "#print(loss.item())"
      ],
      "metadata": {
        "id": "jiKD-PPOdClh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(stepi, lossi)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "ZBAYXZ_rdGIG",
        "outputId": "4da78334-f8d9-416f-9d09-64b66823372a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7fae0cd9ed90>]"
            ]
          },
          "metadata": {},
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# training loss\n",
        "emb = C[Xtr] # (32, 3, 2)\n",
        "h = torch.tanh(emb.view(-1, 30) @ W1 + b1) # (32, 100)\n",
        "logits = h @ W2 + b2 # (32, 27)\n",
        "loss = F.cross_entropy(logits, Ytr)\n",
        "loss"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZlhYinNgdIVF",
        "outputId": "b11e07b9-827b-416d-b70d-a206a6124598"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(2.1335, grad_fn=<NllLossBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# validation loss\n",
        "emb = C[Xdev] # (32, 3, 2)\n",
        "h = torch.tanh(emb.view(-1, 30) @ W1 + b1) # (32, 100)\n",
        "logits = h @ W2 + b2 # (32, 27)\n",
        "loss = F.cross_entropy(logits, Ydev)\n",
        "loss"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cJH3DDu4dLTM",
        "outputId": "d87a5a80-edcf-4fb0-9984-b9dc553521cd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(2.1635, grad_fn=<NllLossBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# test loss\n",
        "emb = C[Xte] # (32, 3, 2)\n",
        "h = torch.tanh(emb.view(-1, 30) @ W1 + b1) # (32, 100)\n",
        "logits = h @ W2 + b2 # (32, 27)\n",
        "loss = F.cross_entropy(logits, Yte)\n",
        "loss"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DwOgCg1vdNhe",
        "outputId": "766db7cb-dbfd-4981-e3c1-93cfa87521f5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(2.1637, grad_fn=<NllLossBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# visualize dimensions 0 and 1 of the embedding matrix C for all characters\n",
        "plt.figure(figsize=(8,8))\n",
        "plt.scatter(C[:,0].data, C[:,1].data, s=200)\n",
        "for i in range(C.shape[0]):\n",
        "    plt.text(C[i,0].item(), C[i,1].item(), itos[i], ha=\"center\", va=\"center\", color='white')\n",
        "plt.grid('minor')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 483
        },
        "id": "CCWLYP9GdPP4",
        "outputId": "6e2e7f56-60a1-4312-e964-7ccbfe5f3a25"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# sample from the model\n",
        "g = torch.Generator().manual_seed(2147483647 + 10)\n",
        "\n",
        "for _ in range(20):\n",
        "\n",
        "    out = []\n",
        "    context = [0] * block_size # initialize with all ...\n",
        "    while True:\n",
        "      emb = C[torch.tensor([context])] # (1,block_size,d)\n",
        "      h = torch.tanh(emb.view(1, -1) @ W1 + b1)\n",
        "      logits = h @ W2 + b2\n",
        "      probs = F.softmax(logits, dim=1)\n",
        "      ix = torch.multinomial(probs, num_samples=1, generator=g).item()\n",
        "      context = context[1:] + [ix]\n",
        "      out.append(ix)\n",
        "      if ix == 0:\n",
        "        break\n",
        "\n",
        "    print(''.join(itos[i] for i in out))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TQUMmgRrdRIA",
        "outputId": "4618124d-0820-424d-99aa-803c84d680c0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "carmah.\n",
            "ambril.\n",
            "khi.\n",
            "mili.\n",
            "thylahnanden.\n",
            "jazhnen.\n",
            "deliah.\n",
            "jareei.\n",
            "ner.\n",
            "kentzeriiv.\n",
            "kaleigh.\n",
            "ham.\n",
            "jory.\n",
            "quint.\n",
            "shon.\n",
            "marian.\n",
            "quinteron.\n",
            "jarynn.\n",
            "jaxen.\n",
            "dus.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "iwENKZS7dT_5"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}