Open oovm opened 2 weeks ago
I'm not sure if this is possible, but I would like to have a sample of the merged .safetensors, so if possible, could you tell me where I can download one of them?
The complete flux model should be 17.2GB:
https://huggingface.co/Comfy-Org/flux1-dev/tree/main?show_file_info=flux1-dev-fp8.safetensors
And the difference from the original folder structure is as follows:
scheduler/ # not mapping
text_encoder/ # -> text_encoders.clip_l.transformer.{KEY}
? # -> text_encoders.clip_l.logit_scale
text_encoder_2/ # -> text_encoders.t5xxl.transformer.{KEY}
? # -> text_encoders.t5xxl.logit_scale
tokenizer/ # no weights
tokenizer_2/ # no weights
transformer/ # not mapping?
vae/ # not mapping?
ae.safetensors # -> vae.{KEY}
flux1-dev.safetensors # -> model.diffusion_model.{KEY}
I'm glad to see that merging clip weights is now supported today.
I wonder if it's possible to merge all weights (clip, t5, unet, vae) into one file for ComfyUI to use.
Each part of the input (clip, t5, unet, vae, lora) is a separate
.safetensors