Fix mixed precision fine-tuning for text-to-image-lora-sdxl example. (huggingface#6751)

* Fix mixed precision fine-tuning for text-to-image-lora-sdxl example.

* fix text_encoder_two bug.

---------

Co-authored-by: Sajad Norouzi <[email protected]>

Author: 2 people

examples/text_to_image/train_text_to_image_lora_sdxl.py

@@ -35,7 +35,7 @@
 from datasets import load_dataset
 from huggingface_hub import create_repo, upload_folder
 from packaging import version
-from peft import LoraConfig
+from peft import LoraConfig, set_peft_model_state_dict
 from peft.utils import get_peft_model_state_dict
 from torchvision import transforms
 from torchvision.transforms.functional import crop
@@ -51,8 +51,13 @@
 )
 from diffusers.loaders import LoraLoaderMixin
 from diffusers.optimization import get_scheduler
-from diffusers.training_utils import cast_training_params, compute_snr
-from diffusers.utils import check_min_version, convert_state_dict_to_diffusers, is_wandb_available
+from diffusers.training_utils import _set_state_dict_into_text_encoder, cast_training_params, compute_snr
+from diffusers.utils import (
+    check_min_version,
+    convert_state_dict_to_diffusers,
+    convert_unet_state_dict_to_peft,
+    is_wandb_available,
+)
 from diffusers.utils.import_utils import is_xformers_available
 from diffusers.utils.torch_utils import is_compiled_module
 
@@ -629,14 +634,6 @@ def main(args):
         text_encoder_one.add_adapter(text_lora_config)
         text_encoder_two.add_adapter(text_lora_config)
 
-    # Make sure the trainable params are in float32.
-    if args.mixed_precision == "fp16":
-        models = [unet]
-        if args.train_text_encoder:
-            models.extend([text_encoder_one, text_encoder_two])
-            # only upcast trainable parameters (LoRA) into fp32
-            cast_training_params(models, dtype=torch.float32)
-
     def unwrap_model(model):
         model = accelerator.unwrap_model(model)
         model = model._orig_mod if is_compiled_module(model) else model
@@ -693,18 +690,34 @@ def load_model_hook(models, input_dir):
             else:
                 raise ValueError(f"unexpected save model: {model.__class__}")
 
-        lora_state_dict, network_alphas = LoraLoaderMixin.lora_state_dict(input_dir)
-        LoraLoaderMixin.load_lora_into_unet(lora_state_dict, network_alphas=network_alphas, unet=unet_)
+        lora_state_dict, _ = LoraLoaderMixin.lora_state_dict(input_dir)
+        unet_state_dict = {f'{k.replace("unet.", "")}': v for k, v in lora_state_dict.items() if k.startswith("unet.")}
+        unet_state_dict = convert_unet_state_dict_to_peft(unet_state_dict)
+        incompatible_keys = set_peft_model_state_dict(unet_, unet_state_dict, adapter_name="default")
+        if incompatible_keys is not None:
+            # check only for unexpected keys
+            unexpected_keys = getattr(incompatible_keys, "unexpected_keys", None)
+            if unexpected_keys:
+                logger.warning(
+                    f"Loading adapter weights from state_dict led to unexpected keys not found in the model: "
+                    f" {unexpected_keys}. "
+                )
 
-        text_encoder_state_dict = {k: v for k, v in lora_state_dict.items() if "text_encoder." in k}
-        LoraLoaderMixin.load_lora_into_text_encoder(
-            text_encoder_state_dict, network_alphas=network_alphas, text_encoder=text_encoder_one_
-        )
+        if args.train_text_encoder:
+            _set_state_dict_into_text_encoder(lora_state_dict, prefix="text_encoder.", text_encoder=text_encoder_one_)
 
-        text_encoder_2_state_dict = {k: v for k, v in lora_state_dict.items() if "text_encoder_2." in k}
-        LoraLoaderMixin.load_lora_into_text_encoder(
-            text_encoder_2_state_dict, network_alphas=network_alphas, text_encoder=text_encoder_two_
-        )
+            _set_state_dict_into_text_encoder(
+                lora_state_dict, prefix="text_encoder_2.", text_encoder=text_encoder_two_
+            )
+
+        # Make sure the trainable params are in float32. This is again needed since the base models
+        # are in `weight_dtype`. More details:
+        # https://github.com/huggingface/diffusers/pull/6514#discussion_r1449796804
+        if args.mixed_precision == "fp16":
+            models = [unet_]
+            if args.train_text_encoder:
+                models.extend([text_encoder_one_, text_encoder_two_])
+            cast_training_params(models, dtype=torch.float32)
 
     accelerator.register_save_state_pre_hook(save_model_hook)
     accelerator.register_load_state_pre_hook(load_model_hook)
@@ -725,6 +738,13 @@ def load_model_hook(models, input_dir):
             args.learning_rate * args.gradient_accumulation_steps * args.train_batch_size * accelerator.num_processes
         )
 
+    # Make sure the trainable params are in float32.
+    if args.mixed_precision == "fp16":
+        models = [unet]
+        if args.train_text_encoder:
+            models.extend([text_encoder_one, text_encoder_two])
+        cast_training_params(models, dtype=torch.float32)
+
     # Use 8-bit Adam for lower memory usage or to fine-tune the model in 16GB GPUs
     if args.use_8bit_adam:
         try: