from absl import app, flags, logging from ml_collections import config_flags from accelerate import Accelerator from accelerate.utils import set_seed from accelerate.logging import get_logger from diffusers import StableDiffusionPipeline, DDIMScheduler from diffusers.loaders import AttnProcsLayers from diffusers.models.attention_processor import LoRAAttnProcessor import ddpo_pytorch.prompts import ddpo_pytorch.rewards from ddpo_pytorch.stat_tracking import PerPromptStatTracker from ddpo_pytorch.diffusers_patch.pipeline_with_logprob import pipeline_with_logprob from ddpo_pytorch.diffusers_patch.ddim_with_logprob import ddim_step_with_logprob import torch import tqdm FLAGS = flags.FLAGS config_flags.DEFINE_config_file("config", "config/base.py", "Training configuration.") logger = get_logger(__name__) def main(_): # basic Accelerate and logging setup config = FLAGS.config accelerator = Accelerator( log_with="all", mixed_precision=config.mixed_precision, project_dir=config.logdir, ) if accelerator.is_main_process: accelerator.init_trackers(project_name="ddpo-pytorch", config=config) logger.info(config) # set seed set_seed(config.seed) # load scheduler, tokenizer and models. pipeline = StableDiffusionPipeline.from_pretrained(config.pretrained.model, revision=config.pretrained.revision) # freeze parameters of models to save more memory pipeline.unet.requires_grad_(False) pipeline.vae.requires_grad_(False) pipeline.text_encoder.requires_grad_(False) # disable safety checker pipeline.safety_checker = None # make the progress bar nicer pipeline.set_progress_bar_config( position=1, disable=not accelerator.is_local_main_process, leave=False, ) # switch to DDIM scheduler pipeline.scheduler = DDIMScheduler.from_config(pipeline.scheduler.config) # For mixed precision training we cast all non-trainable weigths (vae, non-lora text_encoder and non-lora unet) to half-precision # as these weights are only used for inference, keeping weights in full precision is not required. weight_dtype = torch.float32 if accelerator.mixed_precision == "fp16": weight_dtype = torch.float16 elif accelerator.mixed_precision == "bf16": weight_dtype = torch.bfloat16 # Move unet, vae and text_encoder to device and cast to weight_dtype pipeline.unet.to(accelerator.device, dtype=weight_dtype) pipeline.vae.to(accelerator.device, dtype=weight_dtype) pipeline.text_encoder.to(accelerator.device, dtype=weight_dtype) # Set correct lora layers lora_attn_procs = {} for name in pipeline.unet.attn_processors.keys(): cross_attention_dim = None if name.endswith("attn1.processor") else pipeline.unet.config.cross_attention_dim if name.startswith("mid_block"): hidden_size = pipeline.unet.config.block_out_channels[-1] elif name.startswith("up_blocks"): block_id = int(name[len("up_blocks.")]) hidden_size = list(reversed(pipeline.unet.config.block_out_channels))[block_id] elif name.startswith("down_blocks"): block_id = int(name[len("down_blocks.")]) hidden_size = pipeline.unet.config.block_out_channels[block_id] lora_attn_procs[name] = LoRAAttnProcessor(hidden_size=hidden_size, cross_attention_dim=cross_attention_dim) pipeline.unet.set_attn_processor(lora_attn_procs) lora_layers = AttnProcsLayers(pipeline.unet.attn_processors) # Enable TF32 for faster training on Ampere GPUs, # cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices if config.allow_tf32: torch.backends.cuda.matmul.allow_tf32 = True if config.train.scale_lr: config.train.learning_rate = ( config.train.learning_rate * config.train.gradient_accumulation_steps * config.train.batch_size * accelerator.num_processes ) # Initialize the optimizer if config.train.use_8bit_adam: try: import bitsandbytes as bnb except ImportError: raise ImportError( "Please install bitsandbytes to use 8-bit Adam. You can do so by running `pip install bitsandbytes`" ) optimizer_cls = bnb.optim.AdamW8bit else: optimizer_cls = torch.optim.AdamW optimizer = optimizer_cls( lora_layers.parameters(), lr=config.train.learning_rate, betas=(config.train.adam_beta1, config.train.adam_beta2), weight_decay=config.train.adam_weight_decay, eps=config.train.adam_epsilon, ) # prepare prompt and reward fn prompt_fn = getattr(ddpo_pytorch.prompts, config.prompt_fn) reward_fn = getattr(ddpo_pytorch.rewards, config.reward_fn)() # Prepare everything with our `accelerator`. lora_layers, optimizer = accelerator.prepare(lora_layers, optimizer) # Train! samples_per_epoch = config.sample.batch_size * accelerator.num_processes * config.sample.num_batches_per_epoch total_train_batch_size = ( config.train.batch_size * accelerator.num_processes * config.train.gradient_accumulation_steps ) assert config.sample.batch_size % config.train.batch_size == 0 assert samples_per_epoch % total_train_batch_size == 0 logger.info("***** Running training *****") logger.info(f" Num Epochs = {config.num_epochs}") logger.info(f" Sample batch size per device = {config.sample.batch_size}") logger.info(f" Train batch size per device = {config.train.batch_size}") logger.info(f" Gradient Accumulation steps = {config.train.gradient_accumulation_steps}") logger.info("") logger.info(f" Total number of samples per epoch = {samples_per_epoch}") logger.info(f" Total train batch size (w. parallel, distributed & accumulation) = {total_train_batch_size}") logger.info(f" Number of gradient updates per inner epoch = {samples_per_epoch // total_train_batch_size}") logger.info(f" Number of inner epochs = {config.train.num_inner_epochs}") neg_prompt_embed = pipeline.text_encoder( pipeline.tokenizer( [""], return_tensors="pt", padding="max_length", truncation=True, max_length=pipeline.tokenizer.model_max_length, ).input_ids.to(accelerator.device) )[0] sample_neg_prompt_embeds = neg_prompt_embed.repeat(config.sample.batch_size, 1, 1) train_neg_prompt_embeds = neg_prompt_embed.repeat(config.train.batch_size, 1, 1) if config.per_prompt_stat_tracking: stat_tracker = PerPromptStatTracker( config.per_prompt_stat_tracking.buffer_size, config.per_prompt_stat_tracking.min_count, ) for epoch in range(config.num_epochs): #################### SAMPLING #################### samples = [] prompts = [] for i in tqdm.tqdm( range(config.sample.num_batches_per_epoch), desc=f"Epoch {epoch}: sampling", disable=not accelerator.is_local_main_process, position=0, ): # generate prompts prompts, prompt_metadata = zip( *[prompt_fn(**config.prompt_fn_kwargs) for _ in range(config.sample.batch_size)] ) # encode prompts prompt_ids = pipeline.tokenizer( prompts, return_tensors="pt", padding="max_length", truncation=True, max_length=pipeline.tokenizer.model_max_length, ).input_ids.to(accelerator.device) prompt_embeds = pipeline.text_encoder(prompt_ids)[0] # sample pipeline.unet.eval() pipeline.vae.eval() images, _, latents, log_probs = pipeline_with_logprob( pipeline, prompt_embeds=prompt_embeds, negative_prompt_embeds=sample_neg_prompt_embeds, num_inference_steps=config.sample.num_steps, guidance_scale=config.sample.guidance_scale, eta=config.sample.eta, output_type="pt", ) latents = torch.stack(latents, dim=1) # (batch_size, num_steps + 1, 4, 64, 64) log_probs = torch.stack(log_probs, dim=1) # (batch_size, num_steps, 1) timesteps = pipeline.scheduler.timesteps.repeat(config.sample.batch_size, 1) # (batch_size, num_steps) # compute rewards rewards, reward_metadata = reward_fn(images, prompts, prompt_metadata) samples.append( { "prompt_ids": prompt_ids, "prompt_embeds": prompt_embeds, "timesteps": timesteps, "latents": latents[:, :-1], # each entry is the latent before timestep t "next_latents": latents[:, 1:], # each entry is the latent after timestep t "log_probs": log_probs, "rewards": torch.as_tensor(rewards), } ) # collate samples into dict where each entry has shape (num_batches_per_epoch * sample.batch_size, ...) samples = {k: torch.cat([s[k] for s in samples]) for k in samples[0].keys()} # gather rewards across processes rewards = accelerator.gather(samples["rewards"]).cpu().numpy() # per-prompt mean/std tracking if config.per_prompt_stat_tracking: # gather the prompts across processes prompt_ids = accelerator.gather(samples["prompt_ids"]).cpu().numpy() prompts = pipeline.tokenizer.batch_decode(prompt_ids, skip_special_tokens=True) advantages = stat_tracker.update(prompts, rewards) else: advantages = (rewards - rewards.mean()) / (rewards.std() + 1e-8) # ungather advantages; we only need to keep the entries corresponding to the samples on this process samples["advantages"] = ( torch.as_tensor(advantages) .reshape(accelerator.num_processes, -1)[accelerator.process_index] .to(accelerator.device) ) del samples["rewards"] del samples["prompt_ids"] total_batch_size, num_timesteps = samples["timesteps"].shape assert total_batch_size == config.sample.batch_size * config.sample.num_batches_per_epoch assert num_timesteps == config.sample.num_steps #################### TRAINING #################### for inner_epoch in range(config.train.num_inner_epochs): # shuffle samples along batch dimension indices = torch.randperm(total_batch_size, device=accelerator.device) samples = {k: v[indices] for k, v in samples.items()} # shuffle along time dimension, independently for each sample for i in range(total_batch_size): indices = torch.randperm(num_timesteps, device=accelerator.device) for key in ["timesteps", "latents", "next_latents"]: samples[key][i] = samples[key][i][indices] # rebatch for training samples_batched = {k: v.reshape(-1, config.train.batch_size, *v.shape[1:]) for k, v in samples.items()} # dict of lists -> list of dicts for easier iteration samples_batched = [dict(zip(samples_batched, x)) for x in zip(*samples_batched.values())] # train for i, sample in tqdm.tqdm( list(enumerate(samples_batched)), desc=f"Outer epoch {epoch}, inner epoch {inner_epoch}: training", position=0, ): if config.train.cfg: # concat negative prompts to sample prompts to avoid two forward passes embeds = torch.cat([train_neg_prompt_embeds, sample["prompt_embeds"]]) else: embeds = sample["prompt_embeds"] for j in tqdm.trange( num_timesteps, desc=f"Timestep", position=1, leave=False, ): with accelerator.accumulate(pipeline.unet): if config.train.cfg: noise_pred = pipeline.unet( torch.cat([sample["latents"][:, j]] * 2), torch.cat([sample["timesteps"][:, j]] * 2), embeds, ).sample noise_pred_uncond, noise_pred_text = noise_pred.chunk(2) noise_pred = noise_pred_uncond + config.sample.guidance_scale * ( noise_pred_text - noise_pred_uncond ) else: noise_pred = pipeline.unet( sample["latents"][:, j], sample["timesteps"][:, j], embeds ).sample _, log_prob = ddim_step_with_logprob( pipeline.scheduler, noise_pred, sample["timesteps"][:, j], sample["latents"][:, j], eta=config.sample.eta, prev_sample=sample["next_latents"][:, j], ) # ppo logic advantages = torch.clamp( sample["advantages"][:, j], -config.train.adv_clip_max, config.train.adv_clip_max ) ratio = torch.exp(log_prob - sample["log_probs"][:, j]) unclipped_loss = -advantages * ratio clipped_loss = -advantages * torch.clamp( ratio, 1.0 - config.train.clip_range, 1.0 + config.train.clip_range ) loss = torch.mean(torch.maximum(unclipped_loss, clipped_loss)) # debugging values info = {} # John Schulman says that (ratio - 1) - log(ratio) is a better # estimator, but most existing code uses this so... # http://joschu.net/blog/kl-approx.html info["approx_kl"] = 0.5 * torch.mean((log_prob - sample["log_probs"][:, j]) ** 2) info["clipfrac"] = torch.mean(torch.abs(ratio - 1.0) > config.train.clip_range) info["loss"] = loss # backward pass accelerator.backward(loss) if accelerator.sync_gradients: accelerator.clip_grad_norm_(lora_layers.parameters(), config.train.max_grad_norm) optimizer.step() optimizer.zero_grad() if __name__ == "__main__": app.run(main)