Skip to content

skip comfyui format conversion for models which support natively #2464

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
bghira merged 1 commit into from
Jan 21, 2026
Merged

skip comfyui format conversion for models which support natively #2464

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions simpletuner/helpers/models/common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1520,6 +1520,9 @@ def save_lora_weights(self, *args, **kwargs):
adapter_metadata.update(value)

lora_format = normalize_lora_format(getattr(self.config, "lora_format", None))
if lora_format == PEFTLoRAFormat.COMFYUI and getattr(self, "NATIVE_COMFYUI_LORA_SUPPORT", False):
logger.info("Skipping ComfyUI LoRA conversion - model has native ComfyUI support")
lora_format = PEFTLoRAFormat.DIFFUSERS # Treat as diffusers format (no-op)
if lora_format == PEFTLoRAFormat.COMFYUI:
import safetensors.torch
from diffusers.loaders.lora_base import LORA_ADAPTER_METADATA_KEY
Expand Down
1 change: 1 addition & 0 deletions simpletuner/helpers/models/flux2/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,7 @@ class Flux2(ImageModelFoundation):
PREDICTION_TYPE = PredictionTypes.FLOW_MATCHING
MODEL_TYPE = ModelTypes.TRANSFORMER
AUTO_LORA_FORMAT_DETECTION = True
NATIVE_COMFYUI_LORA_SUPPORT = True # Flux2 has native ComfyUI LoRA support, no conversion needed
AUTOENCODER_CLASS = AutoencoderKLFlux2
LATENT_CHANNEL_COUNT = 128 # 32 VAE channels × 4 (2×2 pixel shuffle) = 128 transformer channels
VAE_SCALE_FACTOR = 16 # 8x spatial + 2x pixel shuffle
Expand Down