Transformer_MNIST_Generation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Transformer-MNIST-Generation.ipynb",
      "provenance": [],
      "mount_file_id": "1VDznAIhCNh8Z4DBind3iB9ojYO4ShNhl",
      "authorship_tag": "ABX9TyMOAmRAxWDbyxUQlzNL8lFA",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ZYtheNOOOB/iGPT-MNIST/blob/main/Transformer_MNIST_Generation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Transformer Model for MNIST Image Generation\n",
        "\n",
        "A transformer decoder model based on Image-GPT in hand written digits generation.\n",
        "\n",
        "**Author: Jiaye Zhu, 05/05/2022**\n",
        "\n",
        "Credits:\n",
        "\n",
        "[Generative Pretraining from Pixels (OpenAI Paper)](https://cdn.openai.com/papers/Generative_Pretraining_from_Pixels_V2.pdf)\n",
        "\n",
        "[GregorKobsik / ImageTransformer (Github)](https://github.com/GregorKobsik/ImageTransformer)\n",
        "\n",
        "[teddykoker / image-gpt (Github)](https://github.com/teddykoker/image-gpt)\n",
        "\n",
        "[karpathy / minGPT (Github)](https://github.com/karpathy/minGPT)\n",
        "\n",
        "**A result from previous run:**\n",
        "\n",
        "Trained for 39 epochs with best validation loss 0.2591\n",
        "\n",
        "*   1st row: sampled images from test set\n",
        "*   2nd row: masked sample image\n",
        "*   3rd-4th rows: completed images from sample\n",
        "*   5th-6th rows: generated images from one pixel\n",
        "\n",
        "In completed images, 1/20 is non-digit symbol, 1/20 is close but cannot be recognized as digits.\n",
        "\n",
        "In generated images, 2/20 cannot be recognized as digits.\n",
        "\n",
        "![result.png](https://drive.google.com/uc?export=view&id=1NVzjPNeDlCFAcF2mr_2AvU_61OsaksQz)"
      ],
      "metadata": {
        "id": "YbE8xoj2UT0-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Import & Global Settings"
      ],
      "metadata": {
        "id": "rAOQoU94YX73"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VTLM9KRQS18N",
        "outputId": "15d83f7b-6af5-4728-d561-5af1c0b3a226"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting git+https://github.com/ildoonet/pytorch-gradual-warmup-lr.git\n",
            "  Cloning https://github.com/ildoonet/pytorch-gradual-warmup-lr.git to /tmp/pip-req-build-ozb3hhpr\n",
            "  Running command git clone -q https://github.com/ildoonet/pytorch-gradual-warmup-lr.git /tmp/pip-req-build-ozb3hhpr\n",
            "Thu May  5 23:05:50 2022       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   59C    P8    10W /  70W |      0MiB / 15109MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n",
            "Using device: cuda\n"
          ]
        }
      ],
      "source": [
        "%pip install git+https://github.com/ildoonet/pytorch-gradual-warmup-lr.git\n",
        "\n",
        "import os\n",
        "import numpy as np\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import random\n",
        "from torch.utils.data import DataLoader\n",
        "from torchvision import datasets, transforms\n",
        "from tqdm import tqdm\n",
        "from warmup_scheduler import GradualWarmupScheduler\n",
        "\n",
        "os.environ['OMP_NUM_THREADS'] = '1'\n",
        "gpu_info = !nvidia-smi\n",
        "gpu_info = '\\n'.join(gpu_info)\n",
        "if gpu_info.find('failed') >= 0:\n",
        "  print('Not connected to a GPU')\n",
        "else:\n",
        "  print(gpu_info)\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "  device = torch.device('cuda')\n",
        "else:\n",
        "  device = torch.device('cpu')\n",
        "print('Using device:', device)\n",
        "\n",
        "# dir\n",
        "if not os.path.exists('model/'):\n",
        "  os.mkdir('model/')\n",
        "if not os.path.exists('dataset/'):\n",
        "  os.mkdir('dataset/')\n",
        "\n",
        "# set seed\n",
        "def seed_everything(seed):\n",
        "  random.seed(seed)\n",
        "  np.random.seed(seed)\n",
        "  torch.manual_seed(seed)\n",
        "  torch.cuda.manual_seed(seed)\n",
        "\n",
        "seed_everything(seed=42)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Pixel Value Quantizer\n",
        "\n",
        "Quantize pixel value by K-Means Clustering. Set cluster number to *vocab_size*."
      ],
      "metadata": {
        "id": "U_wUqGCKZr_O"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.cluster import MiniBatchKMeans\n",
        "\n",
        "\n",
        "# Euclidean distance\n",
        "def squared_euclidean_distance(a, b):\n",
        "  b = b.t()\n",
        "  a_sq = torch.sum(torch.square(a), dim=1, keepdim=True)\n",
        "  b_sq = torch.sum(torch.square(b), dim=0, keepdim=True)\n",
        "  ab = torch.linalg.matmul(a, b)\n",
        "  return a_sq - 2 * ab + b_sq\n",
        "\n",
        "\n",
        "class PixelQuantizer:\n",
        "  def __init__(self):\n",
        "    self.centroids = None\n",
        "\n",
        "  def load(self):\n",
        "    self.centroids = np.load('model/centroids.npy')\n",
        "\n",
        "  def fit(self, data, n_clusters=16, batch_size=4096):\n",
        "    k_means = MiniBatchKMeans(n_clusters=n_clusters, random_state=42, batch_size=batch_size)\n",
        "    imgs = np.stack([x.numpy() for x, _ in data]).transpose((0, 2, 3, 1))\n",
        "    pixels = imgs.reshape(-1, imgs.shape[-1])\n",
        "    k_means.fit(pixels)\n",
        "    self.centroids = k_means.cluster_centers_\n",
        "    np.save('model/centroids.npy', self.centroids)\n",
        "\n",
        "  def __call__(self, x, centroids):\n",
        "    x = x.reshape((x.shape[0], 1, 1, -1))    # flatten\n",
        "    b, c, h, w = x.shape\n",
        "    # [B, C, H, W] => [B, H, W, C]\n",
        "    x = x.permute(0, 2, 3, 1).contiguous()\n",
        "    x = x.view(-1, c)               # flatten to pixels\n",
        "    d = squared_euclidean_distance(x, centroids)\n",
        "    x = torch.argmin(d, 1)\n",
        "    x = x.view(b, h, w)\n",
        "    x = x.view(x.shape[0], -1)          # (batch, seq_len)\n",
        "    return x\n",
        "\n",
        "  def unquantize(self, x):\n",
        "    for i in range(x.shape[-1]):\n",
        "      x[i] = self.centroids[int(x[i])]\n",
        "    return x"
      ],
      "metadata": {
        "id": "hMRQ2aBzZkhr"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Transformer Decoder Model Class\n",
        "\n",
        "**Decoder block:** Norm -> Multi-head Attn -> Norm -> Feed-forward (plus residual connections)\n",
        "\n",
        "**Generation (test):**\n",
        "\n",
        "1. Feed sequence into model and get logits.\n",
        "2. Use the last logits as prediction.\n",
        "3. Filter *k* largest logits and apply softmax.\n",
        "4. Sample a token (quantized pixel) from filtered logits by their probabilities.\n",
        "5. Append the token to input sequence.\n",
        "6. Repeat 1-5 for a number of steps.\n",
        "7. Unquantize the sequence and reshape.\n",
        "\n"
      ],
      "metadata": {
        "id": "yqDcon9dbjue"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class TransformerDecoder(nn.Module):\n",
        "  def __init__(self, quantizer, input_size=28, embed_dim=16, num_heads=2, num_layers=8, vocab_size=16, dropout=0.2):\n",
        "    super(TransformerDecoder, self).__init__()\n",
        "    self.centroids = nn.Parameter(torch.from_numpy(quantizer.centroids), requires_grad=False)\n",
        "    self.quantizer = quantizer\n",
        "    self.embed_dim = embed_dim\n",
        "\n",
        "    # embeddings\n",
        "    self.token_embed = nn.Embedding(vocab_size, embed_dim)\n",
        "    self.position_embed = nn.Parameter(torch.zeros(1, input_size ** 2, embed_dim))\n",
        "\n",
        "    # transformer layers\n",
        "    self.layers = nn.Sequential(*[DecoderBlock(embed_dim, num_heads, dropout) for _ in range(num_layers)])\n",
        "\n",
        "    self.ln = nn.LayerNorm(embed_dim)\n",
        "    self.head = nn.Linear(embed_dim, vocab_size)\n",
        "\n",
        "    # loss\n",
        "    self.loss = nn.CrossEntropyLoss()\n",
        "\n",
        "  def forward(self, x):\n",
        "    # embeddings\n",
        "    t_emb = self.token_embed(x)\n",
        "    p_emb = self.position_embed[:, :x.shape[1], :]\n",
        "    x = t_emb + p_emb\n",
        "\n",
        "    # transformer\n",
        "    x = self.layers(x)\n",
        "    # layer norm\n",
        "    x = self.ln(x)\n",
        "\n",
        "    logits = self.head(x)\n",
        "    return logits\n",
        "\n",
        "  @torch.no_grad()\n",
        "  def generate(self, x, step, k=20):\n",
        "    for _ in range(step):\n",
        "      logits = self.forward(x)\n",
        "      # use the last logits\n",
        "      logits = logits[:, -1, :]\n",
        "      # top-k sampling\n",
        "      values, indices = torch.topk(logits.squeeze(0), k=k, dim=-1, largest=True, sorted=True)\n",
        "      probs = torch.softmax(values, dim=-1)\n",
        "      pred = torch.multinomial(probs, 1)[0]\n",
        "      # get pixel (quantized)\n",
        "      pixel = indices[pred].reshape(1, -1)\n",
        "      x = torch.cat((x, pixel), dim=1)\n",
        "    # unquantize\n",
        "    x = self.quantizer.unquantize(x.float().squeeze(0).cpu().numpy())\n",
        "    x = x.reshape((28, 28))\n",
        "    return x\n",
        "\n",
        "\n",
        "class DecoderBlock(nn.Module):\n",
        "  def __init__(self, embed_dim, num_heads, dropout):\n",
        "    super(DecoderBlock, self).__init__()\n",
        "    self.ln1 = nn.LayerNorm(embed_dim)\n",
        "    self.ln2 = nn.LayerNorm(embed_dim)\n",
        "    self.attn = nn.MultiheadAttention(embed_dim, num_heads, batch_first=True)\n",
        "    self.ffnn = nn.Sequential(\n",
        "      nn.Linear(embed_dim, embed_dim * 4),\n",
        "      nn.GELU(),\n",
        "      nn.Linear(embed_dim * 4, embed_dim),\n",
        "      nn.Dropout(dropout),\n",
        "    )\n",
        "\n",
        "  def forward(self, x):\n",
        "    # input shape [batch, seq_len]\n",
        "    # future masking\n",
        "    attn_mask = torch.full(\n",
        "      (x.shape[1], x.shape[1]), True, device=x.device, dtype=torch.bool\n",
        "    )\n",
        "    attn_mask = torch.triu(attn_mask, diagonal=1)\n",
        "\n",
        "    x = self.ln1(x)\n",
        "    # multi-head attention\n",
        "    a, _ = self.attn(x, x, x, attn_mask=attn_mask)\n",
        "    x = x + a\n",
        "    # feed forward\n",
        "    m = self.ffnn(self.ln2(x))\n",
        "    # residual connection\n",
        "    x = x + m\n",
        "    return x"
      ],
      "metadata": {
        "id": "1UdAMW3kbD3B"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Training and Tesing"
      ],
      "metadata": {
        "id": "juID9OJXcOjs"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Config"
      ],
      "metadata": {
        "id": "YLnO9WFVcSJ6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class CFG:\n",
        "  max_epoch = 50\n",
        "  batch_size = 64\n",
        "  learning_rate = 3e-3\n",
        "  lr_decay = 0.5               # lr decay rate\n",
        "  lr_warmup = 2               # lr warmup epoch\n",
        "  img_dim = 28\n",
        "  emb_dim = 128               # hidden dim\n",
        "  num_layers = 8\n",
        "  num_heads = 4\n",
        "  vocab_size = 32              # clustering param\n",
        "  betas = (0.9, 0.95)            # optimizer param\n",
        "  dropout = 0.1\n",
        "  k = 20                   # top-k sampling\n",
        "  train_ratio = 0.9\n",
        "  early_stop = 5\n",
        "  model_path = 'model'"
      ],
      "metadata": {
        "id": "mDluN0TocODd"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Dataset (MNIST)"
      ],
      "metadata": {
        "id": "Cjb-2X6ZZTbH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_set = datasets.MNIST(root='dataset/',\n",
        "              transform=transforms.ToTensor(),\n",
        "              train=True,\n",
        "              download=True)\n",
        "test_set = datasets.MNIST(root='dataset/',\n",
        "              transform=transforms.ToTensor(),\n",
        "              train=False)\n",
        "\n",
        "# split train val sets (9:1)\n",
        "train_size = int(len(train_set) * CFG.train_ratio)\n",
        "train_set, val_set = torch.utils.data.random_split(train_set, [train_size, len(train_set)-train_size])\n",
        "print(f'Train size: {len(train_set)}')\n",
        "print(f'Val size: {len(val_set)}')\n",
        "\n",
        "# Quantize\n",
        "quantizer = PixelQuantizer()\n",
        "quantizer.fit(train_set, n_clusters=CFG.vocab_size)\n",
        "print(f'Clustered vocab: {CFG.vocab_size}')\n",
        "\n",
        "train_set = DataLoader(dataset=train_set, batch_size=CFG.batch_size, shuffle=True)\n",
        "val_set = DataLoader(dataset=val_set, batch_size=CFG.batch_size)\n",
        "test_set = DataLoader(dataset=test_set, batch_size=1)"
      ],
      "metadata": {
        "id": "Wse9sK6BYa2h",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3b98358e-9c07-405a-85d5-8bbf68087fa8"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train size: 54000\n",
            "Val size: 6000\n",
            "Clustered vocab: 32\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Data augmentation"
      ],
      "metadata": {
        "id": "B3sqppmw3DDZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "data_aug = torch.nn.Sequential(\n",
        "  # rotation\n",
        "  transforms.RandomApply(torch.nn.ModuleList([\n",
        "    transforms.RandomRotation((-10, 10), interpolation=transforms.InterpolationMode.BILINEAR),\n",
        "  ]), p=0.2),\n",
        "  # scale\n",
        "  transforms.RandomApply(torch.nn.ModuleList([\n",
        "    transforms.Resize((26, 26)),\n",
        "    transforms.Pad((1, 1, 1, 1), fill=0),\n",
        "    # translate\n",
        "    transforms.RandomApply(torch.nn.ModuleList([\n",
        "      transforms.RandomAffine(degrees=0, translate=(0.15, 0.15)),\n",
        "    ]), p=0.5),\n",
        "  ]), p=0.2),\n",
        ")"
      ],
      "metadata": {
        "id": "X3lCNc3R3KDt"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model, Optimizer & Scheduler"
      ],
      "metadata": {
        "id": "rgOXM1v8cwVe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# model\n",
        "model = TransformerDecoder(quantizer=quantizer,\n",
        "              embed_dim=CFG.emb_dim,\n",
        "              vocab_size=CFG.vocab_size,\n",
        "              num_layers=CFG.num_layers,\n",
        "              num_heads=CFG.num_heads,\n",
        "              dropout=CFG.dropout)\n",
        "model.to(device=device)\n",
        "\n",
        "# optimizer\n",
        "optimizer = torch.optim.Adam(model.parameters(),\n",
        "               lr=CFG.learning_rate,\n",
        "               betas=CFG.betas,\n",
        "               eps=1e-8)\n",
        "# lr scheduler\n",
        "scheduler = torch.optim.lr_scheduler.StepLR(optimizer,\n",
        "                       step_size=10,\n",
        "                       gamma=CFG.lr_decay,\n",
        "                       last_epoch=-1)\n",
        "# lr warmup\n",
        "scheduler_warmup = GradualWarmupScheduler(optimizer, \n",
        "                      multiplier=1, \n",
        "                      total_epoch=len(train_set) * CFG.lr_warmup, \n",
        "                      after_scheduler=scheduler)\n",
        "optimizer.zero_grad()  # avoid warning msg, https://github.com/ildoonet/pytorch-gradual-warmup-lr/issues/8\n",
        "optimizer.step()"
      ],
      "metadata": {
        "id": "mLBpD-WCckgN"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Training\n",
        "\n",
        "Use the generative pretraining method proposed in GPT-2: predict the next token from a given sequence (single direction only).\n",
        "\n",
        "Performance is measured by NLL loss between the true and generated sequences."
      ],
      "metadata": {
        "id": "a4vTaDxnkr2L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def process_batch(x):\n",
        "  # x = data_aug(x)   # it seems that adding this augmentation cannot further reduce val loss\n",
        "  x = x.to(device=device)\n",
        "  # flatten and quantize\n",
        "  x = model.quantizer(x, model.centroids)\n",
        "  # predict the next pixel after sequence\n",
        "  x, y = x[:, :-1], x[:, 1:]\n",
        "  return x, y\n",
        "\n",
        "val_loss = None\n",
        "best_loss = float('inf')\n",
        "warmup_step = 0\n",
        "patience = 0\n",
        "# train loop\n",
        "for e in range(CFG.max_epoch):\n",
        "  with tqdm(train_set, unit=\"batch\") as epoch:\n",
        "    for x, _ in epoch:\n",
        "      epoch.set_description(f\"Epoch {e}\")\n",
        "\n",
        "      x, y = process_batch(x)\n",
        "      # forward\n",
        "      optimizer.zero_grad()\n",
        "      logits = model(x)\n",
        "      # generation loss\n",
        "      loss = model.loss(logits.view(-1, logits.size(-1)), y.contiguous().view(-1))\n",
        "      loss.backward()\n",
        "\n",
        "      if warmup_step < len(train_set) * CFG.lr_warmup:\n",
        "        scheduler_warmup.step(warmup_step)\n",
        "        warmup_step += 1\n",
        "      optimizer.step()\n",
        "\n",
        "      epoch.set_postfix(train_loss=loss.item(), val_loss=val_loss)\n",
        "    scheduler.step()\n",
        "  \n",
        "  # validation\n",
        "  model.eval()\n",
        "  loss = []\n",
        "  with torch.no_grad():\n",
        "    for x, _ in val_set:\n",
        "      x, y = process_batch(x)\n",
        "      logits = model(x)\n",
        "      loss.append(model.loss(logits.view(-1, logits.size(-1)), y.contiguous().view(-1)).item())\n",
        "  val_loss = np.average(loss)\n",
        "  model.train()\n",
        "\n",
        "  # early termination\n",
        "  if val_loss > best_loss:\n",
        "    patience += 1\n",
        "    if patience >= CFG.early_stop:\n",
        "      print(f'Early termination at epoch {e} with best val loss {best_loss}.')\n",
        "      break\n",
        "  else:\n",
        "    patience = 0 if val_loss < best_loss else patience\n",
        "    best_loss = val_loss\n",
        "    # save\n",
        "    torch.save(model.state_dict(), os.path.join(CFG.model_path, 'best_model.pth'))"
      ],
      "metadata": {
        "id": "cUNa6GDlj7U5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d08677b1-bea3-4bb4-9636-d2c447b22708"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Epoch 0:   0%|          | 0/844 [00:00<?, ?batch/s]/usr/local/lib/python3.7/dist-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.\n",
            "  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)\n",
            "Epoch 0: 100%|██████████| 844/844 [10:37<00:00,  1.32batch/s, train_loss=0.406, val_loss=None]\n",
            "Epoch 1: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.339, val_loss=0.404]\n",
            "Epoch 2: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.306, val_loss=0.331]\n",
            "Epoch 3: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.31, val_loss=0.311]\n",
            "Epoch 4: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.284, val_loss=0.299]\n",
            "Epoch 5: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.3, val_loss=0.294]\n",
            "Epoch 6: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.306, val_loss=0.291]\n",
            "Epoch 7: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.288, val_loss=0.287]\n",
            "Epoch 8: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.28, val_loss=0.285]\n",
            "Epoch 9: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.273, val_loss=0.282]\n",
            "Epoch 10: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.268, val_loss=0.281]\n",
            "Epoch 11: 100%|██████████| 844/844 [10:41<00:00,  1.32batch/s, train_loss=0.287, val_loss=0.274]\n",
            "Epoch 12: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.269, val_loss=0.273]\n",
            "Epoch 13: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.273, val_loss=0.272]\n",
            "Epoch 14: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.281, val_loss=0.271]\n",
            "Epoch 15: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.287, val_loss=0.27]\n",
            "Epoch 16: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.264, val_loss=0.269]\n",
            "Epoch 17: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.239, val_loss=0.269]\n",
            "Epoch 18: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.279, val_loss=0.268]\n",
            "Epoch 19: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.27, val_loss=0.267]\n",
            "Epoch 20: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.27, val_loss=0.267]\n",
            "Epoch 21: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.262, val_loss=0.264]\n",
            "Epoch 22: 100%|██████████| 844/844 [10:43<00:00,  1.31batch/s, train_loss=0.268, val_loss=0.263]\n",
            "Epoch 23: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.257, val_loss=0.263]\n",
            "Epoch 24: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.252, val_loss=0.262]\n",
            "Epoch 25: 100%|██████████| 844/844 [10:41<00:00,  1.32batch/s, train_loss=0.247, val_loss=0.262]\n",
            "Epoch 26: 100%|██████████| 844/844 [10:41<00:00,  1.31batch/s, train_loss=0.264, val_loss=0.262]\n",
            "Epoch 27: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.263, val_loss=0.262]\n",
            "Epoch 28: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.263, val_loss=0.262]\n",
            "Epoch 29: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.249, val_loss=0.261]\n",
            "Epoch 30: 100%|██████████| 844/844 [10:41<00:00,  1.31batch/s, train_loss=0.255, val_loss=0.261]\n",
            "Epoch 31: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.266, val_loss=0.26]\n",
            "Epoch 32: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.244, val_loss=0.259]\n",
            "Epoch 33: 100%|██████████| 844/844 [10:41<00:00,  1.31batch/s, train_loss=0.248, val_loss=0.259]\n",
            "Epoch 34: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.267, val_loss=0.259]\n",
            "Epoch 35: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.251, val_loss=0.259]\n",
            "Epoch 36: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.247, val_loss=0.26]\n",
            "Epoch 37: 100%|██████████| 844/844 [10:41<00:00,  1.31batch/s, train_loss=0.232, val_loss=0.259]\n",
            "Epoch 38: 100%|██████████| 844/844 [10:42<00:00,  1.31batch/s, train_loss=0.231, val_loss=0.259]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Early termination at epoch 38 with best val loss 0.2591499741724197.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Testing"
      ],
      "metadata": {
        "id": "SUW5rMgSk_pA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# load model\n",
        "quantizer = PixelQuantizer()\n",
        "quantizer.load()\n",
        "\n",
        "model = TransformerDecoder(quantizer=quantizer,\n",
        "              embed_dim=CFG.emb_dim,\n",
        "              vocab_size=CFG.vocab_size,\n",
        "              num_layers=CFG.num_layers,\n",
        "              num_heads=CFG.num_heads,\n",
        "              dropout=0.0)\n",
        "model.to(device=device)\n",
        "model.load_state_dict(torch.load(os.path.join(CFG.model_path, 'best_model.pth')))\n",
        "model.eval()"
      ],
      "metadata": {
        "id": "wLsm67BAlgIu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b77f6d03-9e45-4686-a159-5e05af6fff86"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TransformerDecoder(\n",
              "  (token_embed): Embedding(32, 128)\n",
              "  (layers): Sequential(\n",
              "    (0): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (1): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (2): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (3): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (4): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (5): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (6): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "    (7): DecoderBlock(\n",
              "      (ln1): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (ln2): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "      (attn): MultiheadAttention(\n",
              "        (out_proj): NonDynamicallyQuantizableLinear(in_features=128, out_features=128, bias=True)\n",
              "      )\n",
              "      (ffnn): Sequential(\n",
              "        (0): Linear(in_features=128, out_features=512, bias=True)\n",
              "        (1): GELU()\n",
              "        (2): Linear(in_features=512, out_features=128, bias=True)\n",
              "        (3): Dropout(p=0.0, inplace=False)\n",
              "      )\n",
              "    )\n",
              "  )\n",
              "  (ln): LayerNorm((128,), eps=1e-05, elementwise_affine=True)\n",
              "  (head): Linear(in_features=128, out_features=32, bias=True)\n",
              "  (loss): CrossEntropyLoss()\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Visualization**"
      ],
      "metadata": {
        "id": "CAYiAgtPlJqn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from matplotlib import pyplot as plt\n",
        "\n",
        "# get one sample per digit\n",
        "def create_samples(dataset):\n",
        "  sampled = []\n",
        "  samples = []\n",
        "  masked = []\n",
        "  for img, label in dataset:\n",
        "    if label not in sampled:\n",
        "      samples.append(img.clone())\n",
        "      img[:, :, (img.shape[-1] // 2):, :] = 0\n",
        "      masked.append(img)\n",
        "      sampled.append(label)\n",
        "      if len(sampled) == 10:\n",
        "        break\n",
        "  return samples, masked\n",
        "\n",
        "# plot sample and generated images\n",
        "def visualize_samples(samples, generates):\n",
        "  imgs = samples + generates\n",
        "  n_imgs = len(imgs)\n",
        "  nrow = n_imgs // 10 + 1\n",
        "  ncol = 10\n",
        "  plt.figure(figsize=(20, 10))\n",
        "  for i in range(n_imgs):\n",
        "    img = imgs[i]\n",
        "    img = img.reshape((28, 28))\n",
        "    plt.subplot(nrow, ncol, i + 1)\n",
        "    plt.imshow(img, cmap='gray')\n",
        "    plt.axis('off')\n",
        "  plt.show()\n",
        "  plt.savefig('result.png')\n",
        "\n",
        "\n",
        "# create sample image\n",
        "samples, masked = create_samples(test_set)\n",
        "generated = []"
      ],
      "metadata": {
        "id": "QT2BGPyekpye"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Generate hand written digits\n",
        "\n",
        "**Complete half-masked images**\n",
        "\n",
        "Mask the lower half of sample images and ask the model to complete them. The model should generate the same or similar digits."
      ],
      "metadata": {
        "id": "S5Ctjx3tl-a5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for _ in tqdm(range(2), desc='Test image completion'):\n",
        "  for s in samples:\n",
        "    # take upper half of image\n",
        "    s = s[:, :, :s.shape[-1] // 2, :].to(device)\n",
        "    sq = model.quantizer(s, model.centroids)\n",
        "    steps = (s.shape[-1] ** 2) // 2\n",
        "    img = model.generate(sq, step=steps, k=CFG.k)\n",
        "    generated.append(img)"
      ],
      "metadata": {
        "id": "ZBGTTS3-mFu_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3565c0bc-8862-40b7-a8e0-c02f28567ba8"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Test image completion: 100%|██████████| 2/2 [00:49<00:00, 24.96s/it]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Generate image from scratch**\n",
        "\n",
        "Use a single pixel valued 0 (black), quantized and ask the model to generate a complete image from this seed. The model should give images with random digits."
      ],
      "metadata": {
        "id": "SDt6b6BGmpQw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for _ in tqdm(range(20), desc='Test generation'):\n",
        "  # use pixel value 0 as seed\n",
        "  s = torch.Tensor([0.]).reshape((1, 1, 1, -1)).to(device)\n",
        "  sq = model.quantizer(s, model.centroids)\n",
        "  steps = CFG.img_dim ** 2 - 1\n",
        "  img = model.generate(sq, step=steps, k=CFG.k)\n",
        "  generated.append(img)"
      ],
      "metadata": {
        "id": "MPCKnC-Dmld2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8486d072-3149-4ef9-b35e-00cf42f00c9d"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Test generation: 100%|██████████| 20/20 [01:37<00:00,  4.88s/it]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Visualize**\n",
        "\n",
        "*   1st row: sampled images from test set\n",
        "*   2nd row: masked sample image\n",
        "*   3rd-4th rows: completed images from sample\n",
        "*   5th-6th rows: generated images from one pixel\n",
        "\n"
      ],
      "metadata": {
        "id": "f4pYsoMvmwF7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "visualize_samples(samples + masked, generated)"
      ],
      "metadata": {
        "id": "QZ0p2QrWnLPg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 389
        },
        "outputId": "1aae5f81-6a2c-4ffa-b4ed-86f6c0856043"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x720 with 60 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    }
  ]
}