vqvae_inference.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# prepare data\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision.datasets import FashionMNIST\n",
    "from torchvision.transforms import Compose, Lambda, Resize, ToTensor\n",
    "from PIL import Image\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "transforms = Compose([Resize((32, 32)), ToTensor()])\n",
    "train_dataset = FashionMNIST(\"fashion_mnist\", train=True, download=True, transform=transforms)\n",
    "train_dataloader = DataLoader(train_dataset, batch_size=192, shuffle=False, num_workers=2, pin_memory=True)\n",
    "\n",
    "first_tensor = next(iter(train_dataloader))[0][0].permute(1,2,0)\n",
    "first_image = Image.fromarray(first_tensor[...,0].numpy()*255)\n",
    "plt.imshow(first_image)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_63709/1411620727.py:9: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "  vqvae.load_state_dict(torch.load(\"vae.pt\",  map_location=device))\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# test vae\n",
    "from diffusion.autoencoders import VQModel\n",
    "import torch\n",
    "import torch.nn.functional as F\n",
    "\n",
    "device = \"cuda:0\"\n",
    "\n",
    "vqvae = VQModel()\n",
    "vqvae.load_state_dict(torch.load(\"vae.pt\",  map_location=device))\n",
    "\n",
    "x_in = next(iter(train_dataloader))[0][0].unsqueeze(dim=0)\n",
    "x_out, _ = vqvae(x_in)\n",
    "\n",
    "x_out_img = Image.fromarray(x_out[0].permute(1,2,0)[...,0].detach().numpy()*255)\n",
    "res_img = Image.fromarray((x_in - x_out)[0].permute(1,2,0)[...,0].detach().numpy()*255)\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(10, 5))\n",
    "axes[0].imshow(x_out_img)\n",
    "axes[1].imshow(res_img)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".conda",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}