Last active
March 30, 2026 21:20
-
-
Save Stella-S-Yan/03830a1610dd64faf163b34ee01ad5fb to your computer and use it in GitHub Desktop.
JAX Implementation of Residual-Quantized Variational Autoencoder (RQ-VAE)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "cells": [ | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "# JAX Implementation of Residual-Quantized Variational Autoencoder (RQ-VAE)\n", | |
| "\n", | |
| "The __Residual-Quantized Variational Autoencoder (RQ-VAE)__ is commonly used for __data compression__, particularly in cases where learning discrete latent variables leads to efficient encoding and reconstruction of data.\n", | |
| "\n", | |
| "Recently, RQ-VAE has also been applied to generate item semantic-IDs for generative recommendation systems.\n", | |
| "\n", | |
| "This repository contains a Flax nnx implementation of RQ-VAE, inspired by the original PyTorch implementations of both [VQ-VAE](https://colab.sandbox.google.com/github/zalandoresearch/pytorch-vq-vae/blob/master/vq-vae.ipynb#scrollTo=yAzAH7FIb7Tf) and [RQ-VAE](https://github.com/roomo7time/simple_rqvae)." | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "from __future__ import print_function\n", | |
| "\n", | |
| "from functools import partial\n", | |
| "\n", | |
| "import os\n", | |
| "import matplotlib.pyplot as plt\n", | |
| "import numpy as np\n", | |
| "\n", | |
| "\n", | |
| "import torch\n", | |
| "from torch.utils.data import DataLoader\n", | |
| "\n", | |
| "import torchvision.datasets as datasets\n", | |
| "import torchvision.transforms as transforms\n", | |
| "from torch.utils.data import Subset, DataLoader\n", | |
| "from torch.utils.tensorboard import SummaryWriter\n", | |
| "from datetime import datetime\n", | |
| "\n", | |
| "import jax\n", | |
| "import jax.numpy as jnp\n", | |
| "from flax import nnx\n", | |
| "import optax\n", | |
| "\n", | |
| "import orbax.checkpoint as ocp\n", | |
| "\n", | |
| "jax.clear_caches()\n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "class Residual(nnx.Module):\n", | |
| " \"\"\"\n", | |
| " Improves feature learnign by adding a residual connection\n", | |
| " Contains \n", | |
| " - a few convolutional layers\n", | |
| " - Skip (residual) connection to ease gradient flow.\n", | |
| " - Non-linearity\n", | |
| " \"\"\"\n", | |
| " def __init__(self, in_channels, num_hiddens, num_residual_hiddens, rngs: nnx.Rngs):\n", | |
| " self._block = nnx.Sequential(\n", | |
| " nnx.relu, # passing a function reference, not calling the function\n", | |
| " nnx.Conv(in_features=in_channels,\n", | |
| " out_features=num_residual_hiddens,\n", | |
| " kernel_size=(3,3),\n", | |
| " strides=(1,1), padding=\"SAME\", \n", | |
| " use_bias=False, rngs=rngs),\n", | |
| " nnx.relu,\n", | |
| " nnx.Conv(in_features=num_residual_hiddens,\n", | |
| " out_features=num_hiddens,\n", | |
| " kernel_size=(1,1), \n", | |
| " strides=(1,1), \n", | |
| " use_bias=False, rngs=rngs)\n", | |
| " )\n", | |
| " \n", | |
| " def __call__(self, x):\n", | |
| " # skip connection\n", | |
| " return x + self._block(x)\n", | |
| " \n", | |
| "\n", | |
| "class ResidualStack(nnx.Module):\n", | |
| " \"\"\"\n", | |
| " A sequence of residual blocks \n", | |
| " \"\"\"\n", | |
| " def __init__(self, in_channels, num_hiddens, num_residual_layers, num_residual_hiddens, rngs: nnx.Rngs):\n", | |
| " self._num_residual_layers = num_residual_layers\n", | |
| " self._layers = [Residual(in_channels, num_hiddens, num_residual_hiddens, rngs) for _ in range(num_residual_hiddens)]\n", | |
| " \n", | |
| " def __call__(self, x):\n", | |
| " for i in range(self._num_residual_layers):\n", | |
| " x = self._layers[i](x)\n", | |
| " return nnx.relu(x)\n", | |
| " \n", | |
| "\n", | |
| "class Encoder(nnx.Module):\n", | |
| " \"\"\" \n", | |
| " Maps the input data (e.g., an image or audio signal) into a latent space.\n", | |
| " Instead of directly producing a continuous latent vector (like in standard VAEs),\n", | |
| " the encoder outputs a sequence of latent representations that will be `quantized` using\n", | |
| " residual vector quantization.\n", | |
| " _conv_x are convolutional layers for feature extraction and downsampling\n", | |
| " \"\"\"\n", | |
| " def __init__(self, in_channels, num_hiddens, num_residual_layers, num_residual_hiddens, rngs: nnx.Rngs):\n", | |
| " self._conv_1 = nnx.Conv(in_features=in_channels,\n", | |
| " out_features=num_hiddens // 2,\n", | |
| " kernel_size=(4, 4),\n", | |
| " strides=(2,2), padding=1, \n", | |
| " rngs=rngs)\n", | |
| " self._conv_2 = nnx.Conv(in_features=num_hiddens // 2,\n", | |
| " out_features=num_hiddens,\n", | |
| " kernel_size=(4, 4), \n", | |
| " strides=(2,2), padding=1, \n", | |
| " rngs=rngs)\n", | |
| " self._conv_3 = nnx.Conv(in_features=num_hiddens,\n", | |
| " out_features=num_hiddens,\n", | |
| " kernel_size=(3,3), \n", | |
| " strides=(1,1), padding=1, \n", | |
| " rngs=rngs)\n", | |
| " self._residual_stack = ResidualStack(in_channels=num_hiddens,\n", | |
| " num_hiddens=num_hiddens,\n", | |
| " num_residual_layers=num_residual_layers,\n", | |
| " num_residual_hiddens=num_residual_hiddens,\n", | |
| " rngs=rngs)\n", | |
| " \n", | |
| " def __call__(self, inputs):\n", | |
| " x = self._conv_1(inputs)\n", | |
| " x = nnx.relu(x)\n", | |
| " \n", | |
| " x = self._conv_2(x)\n", | |
| " x = nnx.relu(x)\n", | |
| " \n", | |
| " x = self._conv_3(x)\n", | |
| " \n", | |
| " x = self._residual_stack(x)\n", | |
| " \n", | |
| " return x\n", | |
| " \n", | |
| " \n", | |
| " \n", | |
| "class Decoder(nnx.Module):\n", | |
| " \"\"\"\n", | |
| " Reconstructs the original input from the quantized latent representations. \n", | |
| " It typically consists of deconvolutional (transposed convolution) layers\n", | |
| " \"\"\"\n", | |
| " def __init__(self, in_channels, num_hiddens, num_residual_layers, num_residual_hiddens, rngs: nnx.Rngs):\n", | |
| " \n", | |
| " self._conv_1 = nnx.Conv(in_features=in_channels,\n", | |
| " out_features=num_hiddens,\n", | |
| " kernel_size=(3, 3),\n", | |
| " strides=(1,1),\n", | |
| " padding=\"SAME\", rngs=rngs)\n", | |
| " self._residual_stack = ResidualStack(in_channels=num_hiddens,\n", | |
| " num_hiddens=num_hiddens,\n", | |
| " num_residual_layers=num_residual_layers,\n", | |
| " num_residual_hiddens=num_residual_hiddens,\n", | |
| " rngs=rngs)\n", | |
| " self._conv_trans_1 = nnx.ConvTranspose(in_features=num_hiddens,\n", | |
| " out_features=num_hiddens // 2,\n", | |
| " kernel_size=(4, 4),\n", | |
| " strides=(2, 2),\n", | |
| " padding=\"SAME\",\n", | |
| " rngs=rngs)\n", | |
| " self._conv_trans_2 = nnx.ConvTranspose(in_features=num_hiddens // 2,\n", | |
| " out_features=3,\n", | |
| " kernel_size=(4, 4),\n", | |
| " strides=(2, 2),\n", | |
| " padding=\"SAME\",\n", | |
| " rngs=rngs)\n", | |
| " \n", | |
| " def __call__(self, inputs):\n", | |
| " x = self._conv_1(inputs)\n", | |
| " x = self._residual_stack(x)\n", | |
| " x = self._conv_trans_1(x)\n", | |
| " x = nnx.relu(x)\n", | |
| " x = self._conv_trans_2(x)\n", | |
| " \n", | |
| " return x\n", | |
| " \n", | |
| "\n", | |
| "class VectorQuantizerEMA(nnx.Embed):\n", | |
| " def __init__(self, num_embeddings: int, embedding_dim: int, rngs: nnx.Rngs, decay=0.99, eps=1e-5):\n", | |
| " super().__init__(num_embeddings=num_embeddings, \n", | |
| " features=embedding_dim, \n", | |
| " embedding_init= jax.nn.initializers.normal(1.0),\n", | |
| " rngs=rngs)\n", | |
| " \n", | |
| " self.decay = decay\n", | |
| " self.eps = eps\n", | |
| " self.num_embeddings = num_embeddings\n", | |
| " self.embedding_dim = embedding_dim\n", | |
| " \n", | |
| " self._ema_cluster_size = nnx.Variable(jnp.zeros(num_embeddings))\n", | |
| " self._ema_w = nnx.Variable(nnx.initializers.normal(1.0)(rngs.ema(), (self.num_embeddings, self.embedding_dim)))\n", | |
| " \n", | |
| " self._decay = decay\n", | |
| " self._eps = eps\n", | |
| " \n", | |
| " def compute_distances(self, inputs):\n", | |
| " codebook = self.embedding.raw_value\n", | |
| " \n", | |
| " mean = jnp.mean(codebook).item()\n", | |
| " std = jnp.std(codebook).item()\n", | |
| " \n", | |
| " distances = (\n", | |
| " jnp.sum(inputs ** 2, axis=1, keepdims=True) # (N, 1)\n", | |
| " + jnp.sum(codebook ** 2, axis=1) # (n_embed,)\n", | |
| " - 2.0 * jnp.matmul(inputs, codebook.T) # (N, n_embed)\n", | |
| " )\n", | |
| " \n", | |
| " return distances\n", | |
| "\n", | |
| "\n", | |
| " def __call__(self, inputs, training: bool = False):\n", | |
| " # Flatten input\n", | |
| " flat_inputs = jnp.reshape(inputs,(-1, self.embedding_dim))\n", | |
| " \n", | |
| " codebook = self.embedding.raw_value\n", | |
| " codebook_ema = self._ema_w.raw_value\n", | |
| " \n", | |
| " # Calculate distances\n", | |
| " distances = (\n", | |
| " jnp.sum(flat_inputs ** 2, axis=1, keepdims=True) # (N, 1)\n", | |
| " + jnp.sum(codebook ** 2, axis=1) # (n_embed,)\n", | |
| " - 2.0 * jnp.matmul(flat_inputs, codebook.T) # (N, n_embed)\n", | |
| " )\n", | |
| " \n", | |
| " # Encoding\n", | |
| " emb_idxs = jnp.argmin(distances, axis=-1) # (N, )\n", | |
| " # print(f\"encoding idx: {jnp.unique(emb_idxs)}\")\n", | |
| " encodings = nnx.one_hot(emb_idxs, num_classes=self.num_embeddings) # (N, n_embed)\n", | |
| " \n", | |
| " # Quantize\n", | |
| " \n", | |
| " # Use one-hot encoding to retrieve the corresponding codebook entry = quantization\n", | |
| " quantized = jnp.matmul(encodings, codebook) # (N, embed_dim) \n", | |
| " \n", | |
| " # Unflatten: reshape to original input shape\n", | |
| " quantized = quantized.reshape(inputs.shape)\n", | |
| " \n", | |
| " # Use EMA to update the embedding vectors \n", | |
| " if training:\n", | |
| " # EMA update\n", | |
| " # Sums all input vectors assigned to each codebook entry.\n", | |
| " dw = jnp.matmul(encodings.T, flat_inputs) # (N, n_embed) * (N, embed_dim) -> (n_embed, embed_dim)\n", | |
| " # Stores the ema of inputs assiged to each codebook vector\n", | |
| " self._ema_w.raw_value = codebook_ema * self.decay + (1.0 - self.decay) * dw\n", | |
| " \n", | |
| " self._ema_cluster_size.raw_value = self._ema_cluster_size.raw_value * self.decay + (1.0 - self.decay) * jnp.sum(encodings, axis=0)\n", | |
| " \n", | |
| " # Update codebook (embeddings)\n", | |
| " # Normalize the cluster size\n", | |
| " \"\"\" \n", | |
| " To avoid the case where clusters with fewer assignments dominate the update, you normalize the cluster sizes. \n", | |
| " Smaller clusters will have their influence reduced by the normalized term, because the ratio\n", | |
| " cluster_size_ema[i] / n will be small compared to larger clusters. Larger cluster will have their influcence \n", | |
| " increased or prserved. Preventing the model from collapsing to a small subset of embeddings\n", | |
| " \"\"\"\n", | |
| " n = jnp.sum(self._ema_cluster_size.raw_value)\n", | |
| " self._ema_cluster_size.raw_value = ((self._ema_cluster_size.raw_value + self.eps) / (n + self.num_embeddings * self.eps) * n )\n", | |
| " \n", | |
| " # Normalize embeddings\n", | |
| " self.embedding.raw_value = self._ema_w.raw_value / jnp.expand_dims(self._ema_cluster_size.raw_value, axis=1)\n", | |
| " \n", | |
| " return quantized, encodings \n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 5, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "def compute_perplexity(encodings):\n", | |
| " avg_probs = jnp.mean(encodings, axis=0) # Compute mean across the first dimension (batch)\n", | |
| " perplexity = jnp.exp(-jnp.sum(avg_probs * jnp.log(avg_probs + 1e-10))) # Compute perplexity\n", | |
| " return perplexity\n", | |
| "\n", | |
| "\n", | |
| "def ste(x, x_hat):\n", | |
| " \"\"\"Straight through estimator\"\"\"\n", | |
| " return x + jax.lax.stop_gradient(x_hat - x)\n", | |
| "\n", | |
| " \n", | |
| "class RQVAE(nnx.Module):\n", | |
| " \"\"\"\n", | |
| " num_hiddens: input size to the residual stack.\n", | |
| " num_residual_layers: how many residual blocks the encoder has.\n", | |
| " num_residual_hiddens: hidden layer size within each residual block\n", | |
| " num_embeddings: number of codebook entries\n", | |
| " embedding_dim: codebook entry dimensionality\n", | |
| " depth: the number of residual quantization steps\n", | |
| " decay: \n", | |
| " commitment_cost: \n", | |
| " \"\"\"\n", | |
| " def __init__(self, inchannels, num_hiddens, num_residual_layers, num_residual_hiddens,\n", | |
| " num_embeddings, embedding_dim, commitment_cost, rngs: nnx.Rngs, decay=0.99, eps=1e-5, depth=1):\n", | |
| " \n", | |
| " self._encoder = Encoder(in_channels= inchannels,\n", | |
| " num_hiddens=num_hiddens,\n", | |
| " num_residual_layers=num_residual_layers,\n", | |
| " num_residual_hiddens=num_residual_hiddens,\n", | |
| " rngs=rngs)\n", | |
| " self._pre_vq_conv = nnx.Conv(in_features=num_hiddens,\n", | |
| " out_features=embedding_dim,\n", | |
| " kernel_size=(1,1), strides=(1,1),\n", | |
| " padding=\"SAME\",\n", | |
| " rngs=rngs)\n", | |
| " \n", | |
| " self._quantizer = VectorQuantizerEMA(num_embeddings=num_embeddings,\n", | |
| " embedding_dim=embedding_dim,\n", | |
| " decay=decay,\n", | |
| " eps=eps,\n", | |
| " rngs=rngs)\n", | |
| " self._decoder = Decoder(in_channels=embedding_dim,\n", | |
| " num_hiddens=num_hiddens,\n", | |
| " num_residual_layers=num_residual_layers,\n", | |
| " num_residual_hiddens=num_residual_hiddens,\n", | |
| " rngs=rngs)\n", | |
| " \n", | |
| " self._commitment_cost = commitment_cost\n", | |
| " self._depth = depth\n", | |
| " \n", | |
| " def __call__(self, inputs, training=False, return_metrics=False):\n", | |
| " x = self._encoder(inputs)\n", | |
| " x = self._pre_vq_conv(x) # additional 1x1 conv (to increase model complexity)\n", | |
| " \n", | |
| " # Start residual quantization\n", | |
| " x_nograd = jax.lax.stop_gradient(x)\n", | |
| " _residue = x_nograd\n", | |
| " x_hat = None # current estimation\n", | |
| " \n", | |
| " # Initialize metrics only if needed\n", | |
| " total_perplexity = 0.0 if return_metrics else None\n", | |
| " total_commitment_loss = 0.0 if return_metrics else None\n", | |
| " \n", | |
| " for d in range(self._depth):\n", | |
| " _quantized, _encodings = self._quantizer(_residue, training) # quantize the residual\n", | |
| " \n", | |
| " if d == 0:\n", | |
| " x_hat = _quantized\n", | |
| " else:\n", | |
| " x_hat += _quantized\n", | |
| " \n", | |
| " if return_metrics:\n", | |
| " total_perplexity += compute_perplexity(_encodings)\n", | |
| " total_commitment_loss += jnp.mean(optax.l2_loss(x_hat, x)) * 0.5\n", | |
| " \n", | |
| " _residue = x_nograd - x_hat\n", | |
| " # End quantization\n", | |
| " \n", | |
| " # Straight-through estimation (STE)\n", | |
| " x_hat_flowing = ste(x, x_hat)\n", | |
| " \n", | |
| " # Decoder\n", | |
| " x_recon = self._decoder(x_hat_flowing)\n", | |
| " \n", | |
| " if return_metrics:\n", | |
| " avg_perplexity = total_perplexity / self._depth\n", | |
| " commitment_loss = self._commitment_cost * total_commitment_loss\n", | |
| " return commitment_loss, x_recon, avg_perplexity\n", | |
| " else:\n", | |
| " return x_recon" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stderr", | |
| "output_type": "stream", | |
| "text": [ | |
| "/usr/local/google/home/stellasyan/miniconda3/envs/rec/lib/python3.11/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.\n", | |
| " self.pid = os.fork()\n", | |
| "/usr/local/google/home/stellasyan/miniconda3/envs/rec/lib/python3.11/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.\n", | |
| " self.pid = os.fork()\n" | |
| ] | |
| }, | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "[epoch 0] | loss: 0.054 \n", | |
| "[epoch 1] | loss: 0.017 \n", | |
| "[epoch 2] | loss: 0.013 \n", | |
| "[epoch 3] | loss: 0.011 \n", | |
| "[epoch 4] | loss: 0.010 \n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "\n", | |
| "\n", | |
| "def show(img):\n", | |
| " npimg = img.numpy()\n", | |
| " fig = plt.imshow(np.transpose(npimg, (1,2,0)), interpolation='nearest')\n", | |
| " fig.axes.get_xaxis().set_visible(False)\n", | |
| " fig.axes.get_yaxis().set_visible(False)\n", | |
| " \n", | |
| "training_data = datasets.CIFAR10(root=\"data\", train=True, download=True,\n", | |
| " transform=transforms.Compose([\n", | |
| " transforms.ToTensor(),\n", | |
| " transforms.Normalize((0.5,0.5,0.5), (1.0,1.0,1.0))\n", | |
| " ]))\n", | |
| "\n", | |
| "validation_data = datasets.CIFAR10(root=\"data\", train=False, download=True,\n", | |
| " transform=transforms.Compose([\n", | |
| " transforms.ToTensor(),\n", | |
| " transforms.Normalize((0.5,0.5,0.5), (1.0,1.0,1.0))\n", | |
| " ]))\n", | |
| "\n", | |
| "data_variance = jnp.var(training_data.data / 255.0)\n", | |
| "\n", | |
| "depth = 4\n", | |
| "num_workers = 1\n", | |
| "\n", | |
| "model_name = 'rqvae_depth-%d'%depth\n", | |
| "now_time = datetime.now().strftime(\"%Y-%m-%d-%H-%M-%S\")\n", | |
| "tb_dir_path = './logs/%s/tb_log/%s'%(model_name, now_time)\n", | |
| "\n", | |
| "batch_size = 256\n", | |
| "num_epochs = 50\n", | |
| "\n", | |
| "num_hiddens = 128\n", | |
| "num_residual_hiddens = 32\n", | |
| "num_residual_layers = 2\n", | |
| "in_channels = 3\n", | |
| "\n", | |
| "embedding_dim = 64\n", | |
| "num_embeddings = 256\n", | |
| "\n", | |
| "commitment_cost = 0.25\n", | |
| "\n", | |
| "decay = 0.99\n", | |
| "eps = 1e-5\n", | |
| "\n", | |
| "learning_rate = 1e-3\n", | |
| "\n", | |
| "training_loader = DataLoader(training_data,\n", | |
| " batch_size=batch_size,\n", | |
| " shuffle=True,\n", | |
| " pin_memory=True,\n", | |
| " num_workers=num_workers)\n", | |
| "\n", | |
| "validation_loader = DataLoader(validation_data,\n", | |
| " batch_size=16,\n", | |
| " shuffle=True,\n", | |
| " pin_memory=True,\n", | |
| " num_workers=num_workers)\n", | |
| "\n", | |
| "\n", | |
| "\n", | |
| "model = RQVAE(in_channels, num_hiddens, num_residual_layers, num_residual_hiddens,\n", | |
| " num_embeddings, embedding_dim, commitment_cost,\n", | |
| " decay=decay, eps=eps, depth=depth, rngs=nnx.Rngs(0))\n", | |
| "\n", | |
| "optimizer = nnx.Optimizer(model, optax.adam(learning_rate))\n", | |
| "\n", | |
| "\n", | |
| "def rqvae_loss(model: RQVAE, x: jax.Array, data_variance: float):\n", | |
| " commitment_loss, x_recon, perplexity = model(x, training=True, return_metrics=True)\n", | |
| " recon_loss = jnp.mean(optax.l2_loss(x_recon, x)) * 0.5 / data_variance\n", | |
| " loss = recon_loss + commitment_loss\n", | |
| " return loss\n", | |
| "\n", | |
| "@nnx.jit\n", | |
| "def train_step(model: RQVAE, optimizer: nnx.Optimizer, x: jax.Array, data_variance: jax.Array) -> jax.Array:\n", | |
| " loss, grads = nnx.value_and_grad(rqvae_loss)( model, x, data_variance)\n", | |
| " optimizer.update(grads)\n", | |
| " return loss\n", | |
| "\n", | |
| "# with torch.autograd.set_detect_anomaly(True):\n", | |
| "for epoch in range(num_epochs):\n", | |
| "\n", | |
| " # for (data, _) in tqdm(training_loader):\n", | |
| " for (data, _) in training_loader:\n", | |
| " data = jnp.array(data)\n", | |
| " data = jnp.transpose(data, axes=(0, 2, 3, 1))\n", | |
| " loss = train_step(model, optimizer, data, data_variance)\n", | |
| " \n", | |
| " print('[epoch %d] | loss: %.3f '%(epoch, loss.item()))\n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 7, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "image/png": "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", | |
| "text/plain": [ | |
| "<Figure size 640x480 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| } | |
| ], | |
| "source": [ | |
| "from torchvision.utils import make_grid\n", | |
| "\n", | |
| "(images_test, _) = next(iter(validation_loader))\n", | |
| "original_image_set = make_grid(images_test.cpu() + 0.5, nrow=4)\n", | |
| "show(original_image_set)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Image Reconstruction" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 8, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stderr", | |
| "output_type": "stream", | |
| "text": [ | |
| "/tmp/ipykernel_74858/419158660.py:10: UserWarning: The given NumPy array is not writable, and PyTorch does not support non-writable tensors. This means writing to this tensor will result in undefined behavior. You may want to copy the array to protect its data or make it writable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at /pytorch/torch/csrc/utils/tensor_numpy.cpp:203.)\n", | |
| " reconstructed_images_test = torch.from_numpy(reconstructed_images_test)\n", | |
| "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers). Got range [-0.04147291..1.0743175].\n" | |
| ] | |
| }, | |
| { | |
| "data": { | |
| "image/png": "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", | |
| "text/plain": [ | |
| "<Figure size 640x480 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| } | |
| ], | |
| "source": [ | |
| "# image reconstruction\n", | |
| "images_test = jnp.array(images_test)\n", | |
| "images_test = jnp.transpose(images_test, (0, 2, 3, 1))\n", | |
| "\n", | |
| "reconstructed_images_test = model(images_test)\n", | |
| "\n", | |
| "# Convert back to torch array\n", | |
| "reconstructed_images_test = jnp.transpose(reconstructed_images_test, (0, 3, 1, 2))\n", | |
| "reconstructed_images_test = jax.device_get(reconstructed_images_test)\n", | |
| "reconstructed_images_test = torch.from_numpy(reconstructed_images_test)\n", | |
| "\n", | |
| "\n", | |
| "recon_image_set = make_grid(reconstructed_images_test + 0.5, nrow=4)\n", | |
| "\n", | |
| "show(recon_image_set)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## t-SNE Visualization of Codebook" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 13, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/plain": [ | |
| "<matplotlib.collections.PathCollection at 0x7fb54392b410>" | |
| ] | |
| }, | |
| "execution_count": 13, | |
| "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": [ | |
| "from sklearn.manifold import TSNE\n", | |
| "\n", | |
| "codebook = jax.device_get(model._quantizer.embedding.raw_value)\n", | |
| "\n", | |
| "proj = TSNE(n_components=2, learning_rate='auto', init='random').fit_transform(codebook)\n", | |
| "\n", | |
| "plt.scatter(proj[:, 0], proj[:, 1], alpha=0.3)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [] | |
| } | |
| ], | |
| "metadata": { | |
| "kernelspec": { | |
| "display_name": "base", | |
| "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 | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment