lora key not loaded

[yanliu-github]

System Information:
OS: POSIX
Python Version: 3.10.12 (main, Jul 5 2023, 18:54:27) [GCC 11.2.0]
Embedded Python: false
PyTorch Version: 2.5.1+cu124
RAM:
Total RAM: 440.71 GB
Free RAM: 52.91 GB
Device:
Name: NVIDIA GeForce RTX 4090 (cuda:0) - Supports cudaMallocAsync
Type: CUDA
VRAM Total: 23.64 GB
VRAM Free: 5.48 GB
Torch VRAM:
Total: 19.25 GB
Free: 1.6 GB

log:

got prompt
llm prompt attention_mask shape: torch.Size([1, 161]), masked tokens: 14
clipL prompt attention_mask shape: torch.Size([1, 77]), masked tokens: 17
model_type FLOW
Scheduler config: FrozenDict([('num_train_timesteps', 1000), ('shift', 9.0), ('reverse', True), ('solver', 'euler'), ('n_tokens', None), ('_use_default_values', ['num_train_timesteps', 'n_tokens'])])
Using accelerate to load and assign model weights to device...
Loading LoRA: hyvideo_FastVideo_LoRA-fp8 with strength: 1.0
lora key not loaded: diffusion_model.double_blocks.0.img_attn.proj.diff_b
lora key not loaded: diffusion_model.double_blocks.0.img_attn.proj.lora_down.weight
lora key not loaded: diffusion_model.double_blocks.0.img_attn.proj.lora_up.weight
lora key not loaded: diffusion_model.double_blocks.0.img_attn.qkv.diff_b
lora key not loaded: diffusion_model.double_blocks.0.img_attn.qkv.lora_down.weight
lora key not loaded: diffusion_model.double_blocks.0.img_attn.qkv.lora_up.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.0.diff_b
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.0.lora_down.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.0.lora_up.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.2.diff_b
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.2.lora_down.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mlp.2.lora_up.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mod.lin.diff_b
lora key not loaded: diffusion_model.double_blocks.0.img_mod.lin.lora_down.weight
lora key not loaded: diffusion_model.double_blocks.0.img_mod.lin.lora_up.weight
..........

workflow image:

What is the issue? How should it be resolved?

The models I downloaded from https://huggingface.co/Kijai/HunyuanVideo_comfy/tree/main.

[Fredd4e]

I get the same issue, not sure what it means

[kijai]

The FastVideo LoRA was extracted using the core ComfyUI LoRA extraction node, so it doesn't work with this wrapper currently. It's slightly worse than using the full FastVideo model though, so that's still preferred way to use it. I extracted it as LoRA mostly to test if adjusting the strength is useful, which it can be...but I don't yet have any conclusive results about that.

[yanliu-github]

The FastVideo LoRA was extracted using the core ComfyUI LoRA extraction node, so it doesn't work with this wrapper currently. It's slightly worse than using the full FastVideo model though, so that's still preferred way to use it. I extracted it as LoRA mostly to test if adjusting the strength is useful, which it can be...but I don't yet have any conclusive results about that.

Okay, I think this is a great project. Will you consider supporting this LoRA in the future?