storytold/FineTrainers-Conditioning
2
0
Fork 0
mirror of https://github.com/storytold/FineTrainers-Conditioning.git synced 2026-10-15 00:04:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor

Browse Source
This commit is contained in:
Aryan
2025-01-11 00:54:03 +01:00
parent 014960fd4c
commit d6821c3da8
10 changed files with 112 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files
finetrainers/args.py
+12 -1
View File
@@ -236,6 +236,7 @@ def parse_arguments() -> Args:
def validate_args(args: Args):
_validate_training_args(args)
_validate_validation_args(args)
@@ -455,7 +456,8 @@ def _add_training_arguments(parser: argparse.ArgumentParser) -> None:
"--training_type",
type=str,
default=None,
help="Type of training to perform. Choose between ['lora','sft']",
choices=["lora", "full-finetune"],
help="Type of training to perform. Choose between ['lora', 'full-finetune']",
)
parser.add_argument("--seed", type=int, default=None, help="A seed for reproducible training.")
parser.add_argument(
@@ -874,6 +876,15 @@ def _map_to_args_type(args: Dict[str, Any]) -> Args:
return result_args
def _validate_training_args(args: Args):
if args.training_type == "lora":
assert args.rank is not None, "Rank is required for LoRA training"
assert args.lora_alpha is not None, "LoRA alpha is required for LoRA training"
assert (
args.target_modules is not None and len(args.target_modules) > 0
), "Target modules are required for LoRA training"
def _validate_validation_args(args: Args):
assert args.validation_prompts is not None, "Validation prompts are required for validation"
if args.validation_images is not None:
finetrainers/cogvideox/__init__.py
+2 -1
View File
@@ -1 +1,2 @@
from .cogvideox_lora import COGVIDEOX_T2V_FULL_FINETUNE_CONFIG, COGVIDEOX_T2V_LORA_CONFIG
from .cogvideox_lora import COGVIDEOX_T2V_LORA_CONFIG
from .full_finetune import COGVIDEOX_T2V_FULL_FINETUNE_CONFIG
finetrainers/cogvideox/cogvideox_lora.py
+1 -15
View File
@@ -312,6 +312,7 @@ def _pad_frames(latents: torch.Tensor, patch_size_t: int):
return latents
# TODO(aryan): refactor into model specs for better re-use
COGVIDEOX_T2V_LORA_CONFIG = {
"pipeline_cls": CogVideoXPipeline,
"load_condition_models": load_condition_models,
@@ -326,18 +327,3 @@ COGVIDEOX_T2V_LORA_CONFIG = {
"forward_pass": forward_pass,
"validation": validation,
}
COGVIDEOX_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": CogVideoXPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"calculate_noisy_latents": calculate_noisy_latents,
"forward_pass": forward_pass,
"validation": validation,
}
finetrainers/cogvideox/full_finetune.py
+32
View File
@@ -0,0 +1,32 @@
from diffusers import CogVideoXPipeline
from .cogvideox_lora import (
calculate_noisy_latents,
collate_fn_t2v,
forward_pass,
initialize_pipeline,
load_condition_models,
load_diffusion_models,
load_latent_models,
post_latent_preparation,
prepare_conditions,
prepare_latents,
validation,
)
# TODO(aryan): refactor into model specs for better re-use
COGVIDEOX_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": CogVideoXPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"calculate_noisy_latents": calculate_noisy_latents,
"forward_pass": forward_pass,
"validation": validation,
}
finetrainers/hunyuan_video/__init__.py
+2 -1
View File
@@ -1 +1,2 @@
from .hunyuan_video_lora import HUNYUAN_VIDEO_T2V_FULL_FINETUNE_CONFIG, HUNYUAN_VIDEO_T2V_LORA_CONFIG
from .full_finetune import HUNYUAN_VIDEO_T2V_FULL_FINETUNE_CONFIG
from .hunyuan_video_lora import HUNYUAN_VIDEO_T2V_LORA_CONFIG
finetrainers/hunyuan_video/full_finetune.py
+30
View File
@@ -0,0 +1,30 @@
from diffusers import HunyuanVideoPipeline
from .hunyuan_video_lora import (
collate_fn_t2v,
forward_pass,
initialize_pipeline,
load_condition_models,
load_diffusion_models,
load_latent_models,
post_latent_preparation,
prepare_conditions,
prepare_latents,
validation,
)
# TODO(aryan): refactor into model specs for better re-use
HUNYUAN_VIDEO_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": HunyuanVideoPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"forward_pass": forward_pass,
"validation": validation,
}
finetrainers/hunyuan_video/hunyuan_video_lora.py
+1 -14
View File
@@ -346,6 +346,7 @@ def _get_clip_prompt_embeds(
return {"pooled_prompt_embeds": prompt_embeds}
# TODO(aryan): refactor into model specs for better re-use
HUNYUAN_VIDEO_T2V_LORA_CONFIG = {
"pipeline_cls": HunyuanVideoPipeline,
"load_condition_models": load_condition_models,
@@ -359,17 +360,3 @@ HUNYUAN_VIDEO_T2V_LORA_CONFIG = {
"forward_pass": forward_pass,
"validation": validation,
}
HUNYUAN_VIDEO_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": HunyuanVideoPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"forward_pass": forward_pass,
"validation": validation,
}
finetrainers/ltx_video/__init__.py
+2 -1
View File
@@ -1 +1,2 @@
from .ltx_video_lora import LTX_VIDEO_T2V_FULL_FINETUNE_CONFIG, LTX_VIDEO_T2V_LORA_CONFIG
from .full_finetune import LTX_VIDEO_T2V_FULL_FINETUNE_CONFIG
from .ltx_video_lora import LTX_VIDEO_T2V_LORA_CONFIG
finetrainers/ltx_video/full_finetune.py
+30
View File
@@ -0,0 +1,30 @@
from diffusers import LTXPipeline
from .ltx_video_lora import (
collate_fn_t2v,
forward_pass,
initialize_pipeline,
load_condition_models,
load_diffusion_models,
load_latent_models,
post_latent_preparation,
prepare_conditions,
prepare_latents,
validation,
)
# TODO(aryan): refactor into model specs for better re-use
LTX_VIDEO_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": LTXPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"forward_pass": forward_pass,
"validation": validation,
}
finetrainers/ltx_video/ltx_video_lora.py
-14
View File
@@ -322,17 +322,3 @@ LTX_VIDEO_T2V_LORA_CONFIG = {
"forward_pass": forward_pass,
"validation": validation,
}
LTX_VIDEO_T2V_FULL_FINETUNE_CONFIG = {
"pipeline_cls": LTXPipeline,
"load_condition_models": load_condition_models,
"load_latent_models": load_latent_models,
"load_diffusion_models": load_diffusion_models,
"initialize_pipeline": initialize_pipeline,
"prepare_conditions": prepare_conditions,
"prepare_latents": prepare_latents,
"post_latent_preparation": post_latent_preparation,
"collate_fn": collate_fn_t2v,
"forward_pass": forward_pass,
"validation": validation,
}