MuseV
MuseV copied to clipboard
MuseV: Infinite-length and High Fidelity Virtual Human Video Generation with Visual Conditioned Parallel Denoising
UnicodeDecodeError: 'gbk' codec can't decode byte 0xb9 in position 191: illegal multibyte sequence
Already download the model. WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.0.1+cu118 with CUDA 1108 (you have 2.0.1+cpu) Python 3.10.11 (you have 3.10.6) Please reinstall xformers...
it looks like the tme branch in controlnet-aux has disappeared, but musev relies on it.
您好,我在生成视频的时候发现有些背景图也会动,比如附件视频中,我的背景图里面的水面会有波纹。我只是想让人物动起来,背景保持静止状态。我的prompt是这样的prompt='''(masterpiece, best quality, highres:1),(1girl, solo:1),(eye blinks:1.0),(head wave:1.3)'' https://github.com/TMElyralab/MuseV/assets/12455127/f241dee2-d327-4d90-a096-dec8170af0a7 '。
之前安装好museV,10G显存跑不起来。。只能生成1秒的视频。。 求助各位大师,除了升级显卡,还有啥办法可以让10G显卡也能稳定运行museV么? 比如修改museV的某些参数,适当降低生成视频的质量
Hi Can someone please explain why my output image is being cut off? What is the best settings to get my character to follow the dwpose animation smoothly? Thank you...
text2video 可以跑 (环境4块显卡,共64G) 1.使用默认显卡,video2video跑不了, 报torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 456.00 MiB (GPU 0; 14.57 GiB total capacity; 13.55 GiB already allocated; 192.75 MiB free; 14.12 GiB reserved...
文档和代码里面controlnet_name可以有多个controlnet名称, 比如 `dwpose,depth`, 那么yaml里面的video_path 如何提供对应的多个视频? 是写成这样吗? ``` video_path: - dwpose.mp4 - depth.mp4 ```
File "/MuseV-main/musev/models/attention_processor.py", line 706, in forward query = self.to_q(hidden_states, scale=scale) File "/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) TypeError: forward() got an unexpected keyword argument 'scale' 请问这个怎么解决?
您好!请问, 已经下载了git clone https://huggingface.co/TMElyralab/MuseV ./checkpoints,,已完整下载 在执行 python app,py 需要下载 diffusion_pytorch_model.safetensors 约为1.45G. 请问这个diffusion_pytorch_model.safetensors 是什么模型的文件呢?,哪里进行使用???,有手动下载的地址吗? 信息如下: 
如果人物占比图像很小,或者在图像的四个角,效果会很差。请问是我操作的问题吗