LightX2V
LightX2V copied to clipboard
ModelTC
Light Video Generation Inference Framework
json内容: ```json { "infer_steps": 4, "target_video_length": 81, "text_len": 512, "target_height": 720, "target_width": 1280, "in_dim": 36, "out_dim": 16, "dim": 5120, "num_channels_latents": 16, "num_layers": 30, "num_heads": 40, "multiple_of": 256, "freq_dim": 256, "self_attn_1_type":...
Hi LightX2V team, Thank you for the great work! I can see Lightning/Distill resources for Wan2.2 T2V/I2V, but nothing Animate-specific. Two quick questions: 1. Do you plan to release a...
### Description Briefly describe the bug you encountered. When running ./scripts/wan22/run_wan22_ti2v_i2v.sh (with model_path/lightx2v path configured) It first reports this error File "/host_home/projects/lightx2v/lightx2v/models/video_encoders/hf/wan/vae_2_2.py", line 755, in encode x[:, :, :1, :,...
what's the different between rank32, rank 64, rank128?
Hello, I build the kernels in RTX5090 with cuda-12.8, and when I call the fucntion `cutlass_scaled_nvfp4_mm_sm120` with bias is not `None`, it gives me a `RuntimeError: Error Internal` I tryed:...
40 steps takes too long to get results (16hours in 128gb ram, 5070TI 16GB)
In the model card, it says "We extended support for the Wan2.1-14B-T2V model and performed a 4-step bidirectional distillation process. The modified code is available at Self-Forcing-Plus.". Does "bidirectional" mean...
这是分别用两个lora跑的结果： 1、wan2.2-lightning： https://github.com/user-attachments/assets/ff518b0f-6034-4108-abbd-64e248321db2 2、wan2.1 lora（高低噪声模型都加lora）： https://github.com/user-attachments/assets/71f5cce4-6821-4c72-b48a-92c0b72274c9 是框架不兼容吗，为啥一个结果那么差，大佬们有没有时间可以看看 config 配置如下：wan_moe_i2v_distill.json { "infer_steps": 4, "target_video_length": 97, "text_len": 512, "target_height": 832, "target_width": 480, "self_attn_1_type": "flash_attn3", "cross_attn_1_type": "flash_attn3", "cross_attn_2_type": "flash_attn3", "seed": 42, "sample_guide_scale": [1.0,...
# 问题描述 我尝试将lightx2v/Wan2.1-T2V-14B-StepDistill-CfgDistill-Lightx2v用项目中提供的converter.py转换为diffusers格式。转化成功了，但是加载模型时提示一些权重为空。 # 使用脚本 脚本如下： `python3 convert.py \ --source /model_zoo/lightx2v-Wan-T2V-14B/distill_model.pt \ --output /model_zoo/lightx2v-Wan-T2V-14B-Diffusers \ --direction forward \ --save_by_block ` # 转换结果 # 报错内容： `Some weights of the model checkpoint...
