Going over the Infinite Scail 2 workflow Models: Model: https://huggingface.co/Comfy-Org/SCAIL-2/tree/main/diffusion_models Sam3.1 in checkpoints: https://huggingface.co/Comfy-Org/sam3.1/blob/main/checkpoints/sam3.1_multiplex_fp16.safetensors Vae & Text encoders - wan_2.1_vae.safetensors: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/vae/wan_2.1_vae.safetensors - umt5_xxl_fp8_e4m3fn_scaled.safetensors: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors - clip_vision_h.safetensors: https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/resolve/main/split_files/clip_vision/clip_vision_h.safetensors 4steps lora: https://huggingface.co/lightx2v/Wan2.2-Distill-Loras/resolve/main/wan2.2_i2v_A14b_low_noise_lora_rank64_lightx2v_4step_1022.safetensors DPO lora: https://huggingface.co/Comfy-Org/SCAIL-2/resolve/main/loras/wan2.1_SCAIL_2_DPO_lora_bf16.safetensors slop twerk lora: https://civitai.red/models/2008663/slop-twerk-wan-22-i2v?modelVersionId=2273467 Motion control pack: https://www.skool.com/ai-tech-and-comfyui-workflows-2226/classroom/bc89a535?md=fd9deacaa2c847f4ab5fbd0592c14735

Hi! Is there any tutorial on how to properly train a character Lora for wan so it works well with motion control + scale 2 workflows? I trained a wan 2.1 (14B) character Lora using 50 high-quality images with a good variety of angles, poses, half body, and full body shots. However, the video quality is still inconsistent. Even when generating at the maximum resolution, the character face, eyes, and hair often become distorted, especially during dance or high motion videos. Interestingly, the Lora performs quite well on selfie-style or low motion videos, where the character identity remains much more stable. Does anyone have recommendations for the optimal training configuration (dataset, captions, repeats, steps, learning rate, etc.) specifically for wan 2.1 (14B) character Loras intended for motion heavy video generation? I'd also appreciate any tutorials or best practices for improving facial consistency in high-motion scenes. Any advice or shared settings would be greatly appreciated. Thanks!