modelscope
Which Wan 2.1 LoRAs are supported? Currently I tried 2 and they make 0 difference
Hi i tried these 2 LoRAs but both makes 0 difference on output :(
Can you verify they work?
https://civitai.com/models/1300505/lego-style-wan21-lora
https://civitai.com/models/1300432/fubuki-wan21-lora
I see it is loaded like this @Artiprocher
@Artiprocher can you provide a working lora so i can test my app? thank you
Loading LoRA models from file: LoRAs\lego_wan_30.safetensors
Using wan_video_text_encoder from models/Wan-AI/Wan2.1-T2V-14B/models_t5_umt5-xxl-enc-bf16.pth.
Using wan_video_dit from ['models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00001-of-00006.safetensors', 'models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00002-of-00006.safetensors', 'models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00003-of-00006.safetensors', 'models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00004-of-00006.safetensors', 'models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00005-of-00006.safetensors', 'models/Wan-AI/Wan2.1-T2V-14B/diffusion_pytorch_model-00006-of-00006.safetensors'].
Using wan_video_vae from models/Wan-AI/Wan2.1-T2V-14B/Wan2.1_VAE.pth.
No wan_video_image_encoder models available.
num_persistent_val 10000000000
[CMD] Model loaded successfully.
[CMD] Generating video 1/1 with prompt: l3g0_5ty13 Lego animation style, a close-up shot of a Lego minifigure with sunglasses and a black leather jacket, sitting inside a Lego sports car. His hands grip the tiny steering wheel, and the dashboard has tiny glowing screens. The background is a blurry city filled with neon Lego signs.
[CMD] Using resolution: width=832 height=480
cfg_scale 6 sigma_shift 6
@FurkanGozukara The two lora are not trained using our framework. This is the reason why they cannot be loaded directly. We updated the lora loading strategy.
The first lora is supported in this commit: https://github.com/modelscope/DiffSynth-Studio/commit/84fb61aaafb14b21e286c41ef8982df45af0a4fc
My internet connection is not stable, and I can't download the second Lora. Can you help me test it?
@FurkanGozukara The two lora are not trained using our framework. This is the reason why they cannot be loaded directly. We updated the lora loading strategy.
The first lora is supported in this commit: 84fb61a
My internet connection is not stable, and I can't download the second Lora. Can you help me test it?
Thank you so much. I will test as soon as possible I am also out of home
@FurkanGozukara The two lora are not trained using our framework. This is the reason why they cannot be loaded directly. We updated the lora loading strategy.
The first lora is supported in this commit: 84fb61a
My internet connection is not stable, and I can't download the second Lora. Can you help me test it?
just tested and 2nd lora worked great
ty so much