{ "cells": [ { "cell_type": "markdown", "id": "aa3d1652", "metadata": {}, "source": [ "# Denoising Diffusion Probabilistic Models with miniai" ] }, { "cell_type": "markdown", "id": "b9060758-4a15-4d74-a8f6-9d187ea91873", "metadata": {}, "source": [ "Now that we written our own barebones training library, let's make some progress towards exploring diffusion model and building Stable Diffusion from scratch.\n", "\n", "We'll start with building and training the model described in the seminal 2020 paper [Denoising Diffusion Probabilistic Models](https://arxiv.org/abs/2006.11239) (DDPM). For more context, while diffusion models were technically invented [back in 2015](https://arxiv.org/abs/1503.03585), diffusion models flew under the radar until this 2020 paper since they were complicated and difficult to train. The 2020 paper introducing DDPMs made some crucial assumptions that significantly simplify the model training and generation processes, as we will see here. Later versions of diffusion models all build upon the same framework introduced in this paper.\n", "\n", "Let's get started and train our own DDPM!" ] }, { "cell_type": "markdown", "id": "e97c4f01", "metadata": {}, "source": [ "## Imports" ] }, { "cell_type": "markdown", "id": "93867089-17d2-40cb-a52c-a088b2352929", "metadata": {}, "source": [ "We'll start with some imports." ] }, { "cell_type": "code", "execution_count": null, "id": "8cb2d883-6db2-4a1f-b699-ff8343df0fc8", "metadata": {}, "outputs": [], "source": [ "import pickle,gzip,math,os,time,shutil,torch,random,logging\n", "import fastcore.all as fc,matplotlib as mpl,numpy as np,matplotlib.pyplot as plt\n", "from collections.abc import Mapping\n", "from pathlib import Path\n", "from operator import attrgetter,itemgetter\n", "from functools import partial\n", "from copy import copy\n", "from contextlib import contextmanager\n", "\n", "from fastcore.foundation import L\n", "import torchvision.transforms.functional as TF,torch.nn.functional as F\n", "from torch import tensor,nn,optim\n", "from torch.utils.data import DataLoader,default_collate\n", "from torch.nn import init\n", "from torch.optim import lr_scheduler\n", "from torcheval.metrics import MulticlassAccuracy\n", "from datasets import load_dataset,load_dataset_builder\n", "\n", "from miniai.datasets import *\n", "from miniai.conv import *\n", "from miniai.learner import *\n", "from miniai.activations import *\n", "from miniai.init import *\n", "from miniai.sgd import *\n", "from miniai.resnet import *\n", "from miniai.augment import *" ] }, { "cell_type": "code", "execution_count": null, "id": "e8273fb3", "metadata": {}, "outputs": [], "source": [ "mpl.rcParams['image.cmap'] = 'gray'\n", "logging.disable(logging.WARNING)" ] }, { "cell_type": "markdown", "id": "33e945bc-26a4-4194-ba12-4cbb7b79e49d", "metadata": {}, "source": [ "## Load the dataset\n", "\n", "We will load the dataset from HuggingFace Hub:" ] }, { "cell_type": "code", "execution_count": null, "id": "99edd708", "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "7932a4c5d09c4e41b105b8ca98bfad00", "version_major": 2, "version_minor": 0 }, "text/plain": [ " 0%| | 0/2 [00:00" ] }, { "cell_type": "markdown", "id": "0cff4bef-3f20-4002-8040-eb014bfbe27b", "metadata": {}, "source": [ "The DDPM U-net is a modification of this with some modern tricks like using attention.\n", "\n", "We will cover how U-nets are created and how modules like attention work in future lessons. For now, we'll import the U-net from the diffusers library:" ] }, { "cell_type": "code", "execution_count": null, "id": "563d47e8", "metadata": {}, "outputs": [], "source": [ "from diffusers import UNet2DModel" ] }, { "cell_type": "code", "execution_count": null, "id": "35a6eed7", "metadata": {}, "outputs": [], "source": [ "model = UNet2DModel(in_channels=1, out_channels=1, block_out_channels=(32, 64, 128, 128))" ] }, { "cell_type": "markdown", "id": "64f26317", "metadata": {}, "source": [ "## Training - easy with a callback!" ] }, { "cell_type": "markdown", "id": "c76c3b3c", "metadata": {}, "source": [ "DDPM is trained quite simply in a few steps:\n", "1. randomly select some timesteps in an iterative noising process.\n", "2. Add noise corresponding to this timestep to the original image. For increasing timesteps, the variance of the noise increases.\n", "3. Pass in this noisy image and the timestep to our model\n", "4. Model is trained with an MSE loss between the model output and the amount of noise added to the image\n", "\n", "\n", "We will implement this in a callback. The callback will randomly select the timestep and create the noisy image before setting up our input and ground truth tensors for the model forward pass and loss calculation.\n", "\n", "After training, we need to sample from this model. This is an iterative denoising process starting from pure noise. We simply keep removing noise predicted by the neural network, but we do it with an expected noise schedule that is reverse of what we saw during training. This is also done in our callback." ] }, { "cell_type": "code", "execution_count": null, "id": "aa916302-00c5-4ec0-ac69-de4dccce755f", "metadata": {}, "outputs": [], "source": [ "class DDPMCB(TrainCB):\n", " order = DeviceCB.order+1\n", " def __init__(self, n_steps, beta_min, beta_max):\n", " super().__init__()\n", " self.n_steps,self.βmin,self.βmax = n_steps,beta_min,beta_max\n", " # variance schedule, linearly increased with timestep\n", " self.β = torch.linspace(self.βmin, self.βmax, self.n_steps)\n", " self.α = 1. - self.β \n", " self.ᾱ = torch.cumprod(self.α, dim=0)\n", " self.σ = self.β.sqrt()\n", "\n", " def predict(self, learn): learn.preds = learn.model(*learn.batch[0]).sample\n", " \n", " def before_batch(self, learn):\n", " device = learn.batch[0].device\n", " ε = torch.randn(learn.batch[0].shape, device=device) # noise, x_T\n", " x0 = learn.batch[0] # original images, x_0\n", " self.ᾱ = self.ᾱ.to(device)\n", " n = x0.shape[0]\n", " # select random timesteps\n", " t = torch.randint(0, self.n_steps, (n,), device=device, dtype=torch.long)\n", " ᾱ_t = self.ᾱ[t].reshape(-1, 1, 1, 1).to(device)\n", " xt = ᾱ_t.sqrt()*x0 + (1-ᾱ_t).sqrt()*ε #noisify the image\n", " # input to our model is noisy image and timestep, ground truth is the noise \n", " learn.batch = ((xt, t), ε)\n", " \n", " @torch.no_grad()\n", " def sample(self, model, sz):\n", " device = next(model.parameters()).device\n", " x_t = torch.randn(sz, device=device)\n", " preds = []\n", " for t in reversed(range(self.n_steps)):\n", " t_batch = torch.full((x_t.shape[0],), t, device=device, dtype=torch.long)\n", " z = (torch.randn(x_t.shape) if t > 0 else torch.zeros(x_t.shape)).to(device)\n", " ᾱ_t1 = self.ᾱ[t-1] if t > 0 else torch.tensor(1)\n", " b̄_t = 1 - self.ᾱ[t]\n", " b̄_t1 = 1 - ᾱ_t1\n", " noise_pred = learn.model(x_t, t_batch).sample\n", " x_0_hat = ((x_t - b̄_t.sqrt() * noise_pred)/self.ᾱ[t].sqrt()).clamp(-1,1)\n", " x0_coeff = ᾱ_t1.sqrt()*(1-self.α[t])/b̄_t\n", " xt_coeff = self.α[t].sqrt()*b̄_t1/b̄_t\n", " x_t = x_0_hat*x0_coeff + x_t*xt_coeff + self.σ[t]*z\n", " preds.append(x_t.cpu())\n", " return preds" ] }, { "cell_type": "markdown", "id": "e6d36b3d-7beb-423e-8c43-6c469983a922", "metadata": {}, "source": [ "Okay now we're ready to train a model!\n", "\n", "Let's create our `Learner`. We'll add our callbacks and train with MSE loss.\n", "\n", "We specify the number of timesteps and the minimum and maximum variance for the DDPM model." ] }, { "cell_type": "code", "execution_count": null, "id": "906dabdb", "metadata": {}, "outputs": [], "source": [ "lr = 4e-3\n", "epochs = 5\n", "tmax = epochs * len(dls.train)\n", "sched = partial(lr_scheduler.OneCycleLR, max_lr=lr, total_steps=tmax)\n", "ddpm_cb = DDPMCB(n_steps=1000, beta_min=0.0001, beta_max=0.02)\n", "cbs = [ddpm_cb, DeviceCB(), ProgressCB(plot=True), MetricsCB(), BatchSchedCB(sched)]\n", "learn = Learner(model, dls, nn.MSELoss(), lr=lr, cbs=cbs, opt_func=optim.Adam)" ] }, { "cell_type": "markdown", "id": "ed349459-a2f9-4f8f-9ea2-40b3ebfb0984", "metadata": {}, "source": [ "Now let's run the fit function:" ] }, { "cell_type": "code", "execution_count": null, "id": "ad242778", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
lossepochtrain
0.0590train
0.0240eval
0.0211train
0.0201eval
0.0192train
0.0172eval
0.0173train
0.0173eval
0.0164train
0.0164eval
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "learn.fit(epochs)" ] }, { "cell_type": "code", "execution_count": null, "id": "023eba4c", "metadata": {}, "outputs": [], "source": [ "mdl_path = Path('models')\n", "mdl_path.mkdir(exist_ok=True)" ] }, { "cell_type": "code", "execution_count": null, "id": "228cf5ab", "metadata": {}, "outputs": [], "source": [ "torch.save(learn.model, mdl_path/'fashion_ddpm.pkl')" ] }, { "cell_type": "code", "execution_count": null, "id": "4ea5de33", "metadata": {}, "outputs": [], "source": [ "learn.model = torch.load(mdl_path/'fashion_ddpm.pkl')" ] }, { "cell_type": "markdown", "id": "8a8c8daf-6645-4923-bf82-78f82adcddfd", "metadata": {}, "source": [ "## Inference" ] }, { "cell_type": "markdown", "id": "166a4cd9-df5a-45a0-8469-b0e699b63ab5", "metadata": {}, "source": [ "Now that we've trained our model, let's generate some images with our model:" ] }, { "cell_type": "code", "execution_count": null, "id": "6e98b94f-38c5-4474-9e49-721201f2a188", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "1000" ] }, "execution_count": null, "metadata": {}, "output_type": "execute_result" } ], "source": [ "set_seed(42)\n", "samples = ddpm_cb.sample(learn.model, (16, 1, 32, 32))\n", "len(samples)" ] }, { "cell_type": "code", "execution_count": null, "id": "ae13040d", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "show_images(-samples[-1], figsize=(5,5))" ] }, { "cell_type": "markdown", "id": "d7ac4018-fd51-43b9-afc4-230fe7ea8be3", "metadata": {}, "source": [ "Let's visualize the sampling process:" ] }, { "cell_type": "code", "execution_count": null, "id": "a3d60054", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Using matplotlib backend: \n" ] }, { "data": { "text/html": [ "" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%matplotlib auto\n", "import matplotlib.animation as animation\n", "from IPython.display import display, HTML\n", "\n", "fig,ax = plt.subplots(figsize=(3,3))\n", "def _show_i(i): return show_image(-samples[i][9], ax=ax, animated=True).get_images()\n", "r = L.range(800,990, 5)+L.range(990,1000)+[999]*10\n", "ims = r.map(_show_i)\n", "\n", "animate = animation.ArtistAnimation(fig, ims, interval=50, blit=True, repeat_delay=3000)\n", "display(HTML(animate.to_html5_video()))" ] }, { "cell_type": "markdown", "id": "cf3860f0-da13-4e23-bff8-a56eddc9db4c", "metadata": {}, "source": [ "Note that I only take the steps between 800 and 1000 since most of the previous steps are actually quite noisy. This is a limitation of the noise schedule used for small images, and papers like [Improved DDPM](https://arxiv.org/abs/2102.09672) suggest other noise schedules for this purpose! (Some potential homework: try out the noise schedule from Improved DDPM and see if it helps.)" ] }, { "cell_type": "code", "execution_count": null, "id": "e71b6eb6", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "python3", "language": "python", "name": "python3" } }, "nbformat": 4, "nbformat_minor": 5 }