lora以外も対応

train_textual_inversion.py

@@ -4,6 +4,7 @@ import gc
 import math
 import os
 import toml
+from multiprocessing import Value
 
 from tqdm import tqdm
 import torch
@@ -71,10 +72,6 @@ imagenet_style_templates_small = [
 ]
 
 
-def collate_fn(examples):
-    return examples[0]
-
-
 def train(args):
     if args.output_name is None:
         args.output_name = args.token_string
@@ -186,6 +183,10 @@ def train(args):
     config_util.blueprint_args_conflict(args,blueprint)
     train_dataset_group = config_util.generate_dataset_group_by_blueprint(blueprint.dataset_group)
 
+    current_epoch = Value('i',0)
+    current_step = Value('i',0)
+    collater = train_util.collater_class(current_epoch,current_step)
+
     # make captions: tokenstring tokenstring1 tokenstring2 ...tokenstringn という文字列に書き換える超乱暴な実装
     if use_template:
         print("use template for training captions. is object: {args.use_object_template}")
@@ -251,7 +252,7 @@ def train(args):
         train_dataset_group,
         batch_size=1,
         shuffle=True,
-        collate_fn=collate_fn,
+        collate_fn=collater,
         num_workers=n_workers,
         persistent_workers=args.persistent_data_loader_workers,
     )
@@ -335,13 +336,14 @@ def train(args):
 
     for epoch in range(num_train_epochs):
         print(f"epoch {epoch+1}/{num_train_epochs}")
-        train_dataset_group.set_current_epoch(epoch + 1)
-        train_dataset_group.set_current_step(global_step)
+        current_epoch.value = epoch+1
 
         text_encoder.train()
 
         loss_total = 0
+
         for step, batch in enumerate(train_dataloader):
+            current_step.value = global_step
             with accelerator.accumulate(text_encoder):
                 with torch.no_grad():
                     if "latents" in batch and batch["latents"] is not None: