Moved common training code into separate module

1 files changed, 55 insertions, 120 deletions

diff --git a/train_ti.py b/train_ti.py
index 9616db6..198cf37 100644
--- a/train_ti.py
+++ b/train_ti.py
@@ -7,7 +7,6 @@ import logging
 import json
 from pathlib import Path
 
-import numpy as np
 import torch
 import torch.nn.functional as F
 import torch.utils.checkpoint
@@ -17,7 +16,6 @@ from accelerate.logging import get_logger
 from accelerate.utils import LoggerType, set_seed
 from diffusers import AutoencoderKL, DDPMScheduler, DPMSolverMultistepScheduler, UNet2DConditionModel
 from diffusers.optimization import get_scheduler, get_cosine_with_hard_restarts_schedule_with_warmup
-from PIL import Image
 from tqdm.auto import tqdm
 from transformers import CLIPTextModel, CLIPTokenizer
 from slugify import slugify
@@ -26,6 +24,7 @@ from common import load_text_embeddings, load_text_embedding
 from pipelines.stable_diffusion.vlpn_stable_diffusion import VlpnStableDiffusion
 from data.csv import CSVDataModule, CSVDataItem
 from training.optimization import get_one_cycle_schedule
+from training.util import AverageMeter, CheckpointerBase, freeze_params, save_args
 from models.clip.prompt import PromptProcessor
 
 logger = get_logger(__name__)
@@ -138,7 +137,7 @@ def parse_args():
     parser.add_argument(
         "--tag_dropout",
         type=float,
-        default=0,
+        default=0.1,
         help="Tag dropout probability.",
     )
     parser.add_argument(
@@ -355,27 +354,7 @@ def parse_args():
     return args
 
 
-def freeze_params(params):
-    for param in params:
-        param.requires_grad = False
-
-
-def save_args(basepath: Path, args, extra={}):
-    info = {"args": vars(args)}
-    info["args"].update(extra)
-    with open(basepath.joinpath("args.json"), "w") as f:
-        json.dump(info, f, indent=4)
-
-
-def make_grid(images, rows, cols):
-    w, h = images[0].size
-    grid = Image.new('RGB', size=(cols*w, rows*h))
-    for i, image in enumerate(images):
-        grid.paste(image, box=(i % cols*w, i//cols*h))
-    return grid
-
-
-class Checkpointer:
+class Checkpointer(CheckpointerBase):
     def __init__(
         self,
         datamodule,
@@ -394,21 +373,24 @@ class Checkpointer:
         sample_batch_size,
         seed
     ):
-        self.datamodule = datamodule
+        super().__init__(
+            datamodule=datamodule,
+            output_dir=output_dir,
+            instance_identifier=instance_identifier,
+            placeholder_token=placeholder_token,
+            placeholder_token_id=placeholder_token_id,
+            sample_image_size=sample_image_size,
+            seed=seed or torch.random.seed(),
+            sample_batches=sample_batches,
+            sample_batch_size=sample_batch_size
+        )
+
         self.accelerator = accelerator
         self.vae = vae
         self.unet = unet
         self.tokenizer = tokenizer
         self.text_encoder = text_encoder
         self.scheduler = scheduler
-        self.instance_identifier = instance_identifier
-        self.placeholder_token = placeholder_token
-        self.placeholder_token_id = placeholder_token_id
-        self.output_dir = output_dir
-        self.sample_image_size = sample_image_size
-        self.seed = seed or torch.random.seed()
-        self.sample_batches = sample_batches
-        self.sample_batch_size = sample_batch_size
 
     @torch.no_grad()
     def checkpoint(self, step, postfix):
@@ -431,9 +413,7 @@ class Checkpointer:
         del learned_embeds
 
     @torch.no_grad()
-    def save_samples(self, step, height, width, guidance_scale, eta, num_inference_steps):
-        samples_path = Path(self.output_dir).joinpath("samples")
-
+    def save_samples(self, step, num_inference_steps, guidance_scale=7.5, eta=0.0):
         text_encoder = self.accelerator.unwrap_model(self.text_encoder)
 
         # Save a sample image
@@ -446,71 +426,10 @@ class Checkpointer:
         ).to(self.accelerator.device)
         pipeline.set_progress_bar_config(dynamic_ncols=True)
 
-        train_data = self.datamodule.train_dataloader()
-        val_data = self.datamodule.val_dataloader()
-
-        generator = torch.Generator(device=pipeline.device).manual_seed(self.seed)
-        stable_latents = torch.randn(
-            (self.sample_batch_size, pipeline.unet.in_channels, height // 8, width // 8),
-            device=pipeline.device,
-            generator=generator,
-        )
-
-        with torch.autocast("cuda"), torch.inference_mode():
-            for pool, data, latents in [("stable", val_data, stable_latents), ("val", val_data, None), ("train", train_data, None)]:
-                all_samples = []
-                file_path = samples_path.joinpath(pool, f"step_{step}.jpg")
-                file_path.parent.mkdir(parents=True, exist_ok=True)
-
-                data_enum = enumerate(data)
-
-                batches = [
-                    batch
-                    for j, batch in data_enum
-                    if j * data.batch_size < self.sample_batch_size * self.sample_batches
-                ]
-                prompts = [
-                    prompt.format(identifier=self.instance_identifier)
-                    for batch in batches
-                    for prompt in batch["prompts"]
-                ]
-                nprompts = [
-                    prompt
-                    for batch in batches
-                    for prompt in batch["nprompts"]
-                ]
-
-                for i in range(self.sample_batches):
-                    prompt = prompts[i * self.sample_batch_size:(i + 1) * self.sample_batch_size]
-                    nprompt = nprompts[i * self.sample_batch_size:(i + 1) * self.sample_batch_size]
-
-                    samples = pipeline(
-                        prompt=prompt,
-                        negative_prompt=nprompt,
-                        height=self.sample_image_size,
-                        width=self.sample_image_size,
-                        image=latents[:len(prompt)] if latents is not None else None,
-                        generator=generator if latents is not None else None,
-                        guidance_scale=guidance_scale,
-                        eta=eta,
-                        num_inference_steps=num_inference_steps,
-                        output_type='pil'
-                    ).images
-
-                    all_samples += samples
-
-                    del samples
-
-                image_grid = make_grid(all_samples, self.sample_batches, self.sample_batch_size)
-                image_grid.save(file_path, quality=85)
-
-                del all_samples
-                del image_grid
+        super().save_samples(pipeline, step, num_inference_steps, guidance_scale, eta)
 
         del text_encoder
         del pipeline
-        del generator
-        del stable_latents
 
         if torch.cuda.is_available():
             torch.cuda.empty_cache()
@@ -814,7 +733,14 @@ def main():
     # Only show the progress bar once on each machine.
 
     global_step = 0
-    min_val_loss = np.inf
+
+    avg_loss = AverageMeter()
+    avg_acc = AverageMeter()
+
+    avg_loss_val = AverageMeter()
+    avg_acc_val = AverageMeter()
+
+    max_acc_val = 0.0
 
     checkpointer = Checkpointer(
         datamodule=datamodule,
@@ -835,9 +761,7 @@ def main():
     )
 
     if accelerator.is_main_process:
-        checkpointer.save_samples(
-            0,
-            args.resolution, args.resolution, 7.5, 0.0, args.sample_steps)
+        checkpointer.save_samples(global_step_offset, args.sample_steps)
 
     local_progress_bar = tqdm(
         range(num_update_steps_per_epoch + num_val_steps_per_epoch),
@@ -910,6 +834,8 @@ def main():
                     else:
                         loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean")
 
+                    acc = (model_pred == latents).float().mean()
+
                     accelerator.backward(loss)
 
                     optimizer.step()
@@ -922,8 +848,8 @@ def main():
                         text_encoder.get_input_embeddings(
                         ).weight[index_fixed_tokens] = original_token_embeds[index_fixed_tokens]
 
-                    loss = loss.detach().item()
-                    train_loss += loss
+                    avg_loss.update(loss.detach_(), bsz)
+                    avg_acc.update(acc.detach_(), bsz)
 
                 # Checks if the accelerator has performed an optimization step behind the scenes
                 if accelerator.sync_gradients:
@@ -932,7 +858,13 @@ def main():
 
                     global_step += 1
 
-                logs = {"train/loss": loss, "lr": lr_scheduler.get_last_lr()[0]}
+                logs = {
+                    "train/loss": avg_loss.avg.item(),
+                    "train/acc": avg_acc.avg.item(),
+                    "train/cur_loss": loss.item(),
+                    "train/cur_acc": acc.item(),
+                    "lr": lr_scheduler.get_last_lr()[0],
+                }
 
                 accelerator.log(logs, step=global_step)
 
@@ -941,12 +873,9 @@ def main():
                 if global_step >= args.max_train_steps:
                     break
 
-            train_loss /= len(train_dataloader)
-
             accelerator.wait_for_everyone()
 
             text_encoder.eval()
-            val_loss = 0.0
 
             with torch.inference_mode():
                 for step, batch in enumerate(val_dataloader):
@@ -976,29 +905,37 @@ def main():
 
                     loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean")
 
-                    loss = loss.detach().item()
-                    val_loss += loss
+                    acc = (model_pred == latents).float().mean()
+
+                    avg_loss_val.update(loss.detach_(), bsz)
+                    avg_acc_val.update(acc.detach_(), bsz)
 
                     if accelerator.sync_gradients:
                         local_progress_bar.update(1)
                         global_progress_bar.update(1)
 
-                    logs = {"val/loss": loss}
+                    logs = {
+                        "val/loss": avg_loss_val.avg.item(),
+                        "val/acc": avg_acc_val.avg.item(),
+                        "val/cur_loss": loss.item(),
+                        "val/cur_acc": acc.item(),
+                    }
                     local_progress_bar.set_postfix(**logs)
 
-            val_loss /= len(val_dataloader)
-
-            accelerator.log({"val/loss": val_loss}, step=global_step)
+            accelerator.log({
+                "val/loss": avg_loss_val.avg.item(),
+                "val/acc": avg_acc_val.avg.item(),
+            }, step=global_step)
 
             local_progress_bar.clear()
             global_progress_bar.clear()
 
             if accelerator.is_main_process:
-                if min_val_loss > val_loss:
+                if avg_acc_val.avg.item() > max_acc_val:
                     accelerator.print(
-                        f"Global step {global_step}: Validation loss reached new minimum: {min_val_loss:.2e} -> {val_loss:.2e}")
+                        f"Global step {global_step}: Validation accuracy reached new maximum: {max_acc_val:.2e} -> {avg_acc_val.avg.item():.2e}")
                     checkpointer.checkpoint(global_step + global_step_offset, "milestone")
-                    min_val_loss = val_loss
+                    max_acc_val = avg_acc_val.avg.item()
 
                 if (epoch + 1) % args.checkpoint_frequency == 0:
                     checkpointer.checkpoint(global_step + global_step_offset, "training")
@@ -1007,9 +944,7 @@ def main():
                     })
 
                 if (epoch + 1) % args.sample_frequency == 0:
-                    checkpointer.save_samples(
-                        global_step + global_step_offset,
-                        args.resolution, args.resolution, 7.5, 0.0, args.sample_steps)
+                    checkpointer.save_samples(global_step + global_step_offset, args.sample_steps)
 
         # Create the pipeline using using the trained modules and save it.
         if accelerator.is_main_process: