solution-3-nflows.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# nflows package"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Classes implemented in the previous task\n",
    "\n",
    "from nflows.transforms import AffineCouplingTransform\n",
    "from nflows.transforms.base import CompositeTransform\n",
    "from nflows.flows.base import Flow\n",
    "from nflows.distributions.normal import StandardNormal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Let us try more expressive transforms!\n",
    "\n",
    "from nflows.transforms.autoregressive import MaskedAffineAutoregressiveTransform   # MAF (from the lecture)\n",
    "from nflows.transforms import PiecewiseRationalQuadraticCouplingTransform  # neural spline flows https://arxiv.org/abs/1906.04032\n",
    "from nflows.transforms import BatchNorm  # introduced along with the Real-NVP"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define the flow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In nflows, the scale and shift neural networks are merged into one - this enables a flexible API where any invertible parameterized transform $g_{\\theta}(x)$ is defined by a number or respective parameters $\\theta$. In the case of real-nvp, there are two parameters: $g_{s, t}(x) = x \\cdot \\exp(s) + t$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch import nn, Tensor\n",
    "import torch.nn.functional as F\n",
    "\n",
    "from nflows.nn.nets import ResidualNet\n",
    "\n",
    "def get_nn_for_transform_parameters(context_features: int,  hidden_features: int = 64, num_blocks: int = 2, activation=F.gelu, use_batch_norm: bool = True):\n",
    "    def func(in_dim: int, out_dim: int):\n",
    "        return ResidualNet(\n",
    "            in_dim, out_dim, hidden_features=hidden_features,\n",
    "            context_features=context_features, num_blocks=num_blocks, activation=activation, use_batch_norm=use_batch_norm\n",
    "        )\n",
    "    return func"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of parameters: 7920720\n"
     ]
    }
   ],
   "source": [
    "num_transforms = 10\n",
    "D = 784\n",
    "hid_dim = 64\n",
    "context_features = 2  # number of digits\n",
    "hidden_features = 64\n",
    "num_blocks = 3\n",
    "\n",
    "\n",
    "base_dist = StandardNormal(shape=[D])\n",
    "\n",
    "transforms = []\n",
    "\n",
    "for i in range(num_transforms):\n",
    "    mask = torch.arange(D) % 2 == i % 2\n",
    "\n",
    "    # use random permutation\n",
    "    mask = mask[torch.randperm(D)]\n",
    "\n",
    "    transforms.append(\n",
    "        PiecewiseRationalQuadraticCouplingTransform(mask, get_nn_for_transform_parameters(\n",
    "            context_features=context_features,\n",
    "            hidden_features=hidden_features,\n",
    "            num_blocks=num_blocks,\n",
    "        ), tail_bound=5, tails='linear',\n",
    "        )   \n",
    "    )\n",
    "\n",
    "    transforms.append(BatchNorm(features=D))\n",
    "\n",
    "transform = CompositeTransform(transforms)\n",
    "\n",
    "flow = Flow(transform, base_dist)\n",
    "\n",
    "print('Number of parameters:', sum(p.numel() for p in flow.parameters()))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MNIST dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.optim import Adam, AdamW\n",
    "from torch.utils.data import DataLoader, Subset\n",
    "from torchvision import datasets, transforms\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "# Define the MNIST dataset with preprocessing\n",
    "class NoisyMNIST(datasets.MNIST):\n",
    "    def __init__(self, *args, noise_std=0.1, subset_digits=None, **kwargs):\n",
    "        super().__init__(*args, **kwargs)\n",
    "        self.noise_std = noise_std\n",
    "        self.subset_digits = subset_digits if subset_digits is not None else list(range(10))\n",
    "        self.num_classes = len(self.subset_digits)\n",
    "        self.mapping = {digit: idx for idx, digit in enumerate(self.subset_digits)}\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        img, target = super().__getitem__(index)\n",
    "        noise = torch.randn_like(img) * self.noise_std\n",
    "        img = img + noise\n",
    "        img = (img - img.mean()) / img.std()  # Standardize the image\n",
    "        mapped_target = torch.tensor(self.mapping.get(target, -1))\n",
    "\n",
    "        if mapped_target.item() < 0:\n",
    "            return img, target, None\n",
    "\n",
    "        one_hot_target = F.one_hot(mapped_target, self.num_classes).float()\n",
    "        return img, target, one_hot_target"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pick your two favorite digits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the subset of digits to use\n",
    "subset_digits = [0, 7]  # Modify this list to select the desired digits\n",
    "\n",
    "\n",
    "# Load the MNIST dataset\n",
    "transform = transforms.Compose([transforms.ToTensor()])\n",
    "train_dataset = NoisyMNIST(root='./data', train=True, download=True, transform=transform, subset_digits=subset_digits)\n",
    "test_dataset = NoisyMNIST(root='./data', train=False, download=True, transform=transform, subset_digits=subset_digits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Filter the dataset to include only the desired digits\n",
    "def filter_digits(dataset, digits):\n",
    "    indices = [i for i, (_, target, _) in enumerate(dataset) if target in digits]\n",
    "    return Subset(dataset, indices)\n",
    "\n",
    "\n",
    "train_dataset = filter_digits(train_dataset, subset_digits)\n",
    "test_dataset = filter_digits(test_dataset, subset_digits)\n",
    "\n",
    "train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)\n",
    "test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm.notebook import trange, tqdm\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d253154bffed49ab82690635b06140c8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/10 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8adebdac2b9e44a78a7f58f9904a5929",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "90caf8770506444b9362ffa9f89785e5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c7b97193744d4d6980d9415d5b065f3e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3f08ee8f7b4a4c8cbe8b0414b69d67e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1b9a5d497ba74be1851d7cfdb389cce9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "61bb1f88b6654903afebe1841a002f4e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6112403aea674116813570ea07b7ee5a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "51bfb9d9265841f1a42cb22346006eb3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fa1099afdce64b6eb9699620f3c52283",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9da90cf024f249f2bab8477d73e21230",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/191 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "device = 'cpu'\n",
    "\n",
    "flow.to(device)\n",
    "\n",
    "optim = AdamW(flow.parameters(), lr=1e-3)\n",
    "\n",
    "# Training loop\n",
    "num_epochs = 10\n",
    "\n",
    "pbar = trange(num_epochs)\n",
    "\n",
    "losses = []\n",
    "\n",
    "for epoch in pbar:\n",
    "    flow.train()\n",
    "    total_loss = 0\n",
    "\n",
    "    epoch_pbar = tqdm(train_loader)\n",
    "\n",
    "    for i, batch in enumerate(epoch_pbar):\n",
    "        optim.zero_grad()\n",
    "        imgs, _, context = batch\n",
    "        imgs = imgs.view(imgs.size(0), -1)  # Flatten the images\n",
    "        loss = - flow.log_prob(imgs.to(device), context=context.to(device)).mean() \n",
    "        loss.backward()\n",
    "        optim.step()\n",
    "        total_loss += loss.item() \n",
    "\n",
    "        epoch_pbar.set_description(f'current loss: {total_loss/  (i + 1):.2e}')\n",
    "    \n",
    "    avg_loss = total_loss / len(train_loader)\n",
    "    pbar.set_description(f'Epoch {epoch + 1}/{num_epochs}, Loss: {avg_loss:.2e}')\n",
    "\n",
    "    losses.append(avg_loss)\n",
    "\n",
    "    flow.eval();\n",
    "\n",
    "    with torch.no_grad():\n",
    "        sample = flow.sample(1, context=torch.tensor([[1, 0]]).float().to(device))\n",
    "        img = sample.view(28, 28).cpu().numpy()\n",
    "        plt.imshow(img)\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1aee9a33c50>"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "flow.eval();\n",
    "\n",
    "with torch.no_grad():\n",
    "    img1 = flow.sample(1, context=torch.tensor([[1, 0]]).float().cuda()).view(28, 28).cpu().numpy()\n",
    "    img2 = flow.sample(1, context=torch.tensor([[0, 1]]).float().cuda()).view(28, 28).cpu().numpy()\n",
    "\n",
    "fig, axs = plt.subplots(ncols=2, )\n",
    "axs[0].imshow(img1)\n",
    "axs[1].imshow(img2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}