Colab notebook throwing error while running inference.py about video-retalking HOT 23 OPEN

有解决方案吗？

path：./third_paet/face3d/extract_kp_videos.py
第16行将face_alignment.LandmarksType._2D改为face_alignment.LandmarksType.TWO_D试试

from video-retalking.

Ok, my China friends. The problem been bc AdrianBulat change a model in facealignment so we need to change code in a proj ect. The solution is to change a line 16 to self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType.TWO_D) in /content/video-retalking/third_part/face3d/extract_kp_videos.py from LandmarksType._2D . Then same in line 152 in utils\alignment_stit.py .Have a nice day everyone.

from video-retalking.

from video-retalking.

U need at least CUDA 11.7 torch version. That current that u installed dosent support your awesome card/

from video-retalking.

您至少需要 CUDA 11.7 火炬版本。你安装的电流支持你的真棒卡/

Thanks, according to your hint. My project also ran successfully!! This is so cool!!

from video-retalking.

Hi guys, this is casued by the update of face-alignment, as discussed in 1adrianb/face-alignment#341.

I changed the version of face-alignment from 1.3.5 to 1.3.4 in the requirements.txt.
The colab notebook should now work fine.

from video-retalking.

+1

from video-retalking.

+1

from video-retalking.

有解决方案吗？

from video-retalking.

我将

我将他改为TWO_D后又出现新的错误，我打算也重新配置下环境。可以给一份你anaconda环境中的requirements.txt吗？

from video-retalking.

+1
原来都是正常的，新加其他视频后就出这样的问题？有解决办法吗？

from video-retalking.

解决了吗？

from video-retalking.

我将

我将他改为TWO_D后又出现新的错误，我打算也重新配置下环境。可以给一份你anaconda环境中的requirements.txt吗？

basicsr==1.4.2
kornia==0.5.1
face-alignment==1.3.5
ninja==1.10.2.3
einops==0.4.1
facexlib==0.2.5
librosa==0.9.2
dlib==19.24.0
gradio>=3.7.0
opencv-contrib-python
opencv-python
scikit-image
numpy==1.23.1

from video-retalking.

好的，谢谢！我再去试试。我尝试过anaconda重新安装环境，也试过降低numpy版本一致不行。

from video-retalking.

Same issue, that something wrong with new version of some package, bc initial one worked fine.

from video-retalking.

I started locally and tried to reinstall the environment. Finally, running the project with colad is the same problem. Do you have a solution on your end?

from video-retalking.

Ok, my China friends. The problem been bc AdrianBulat change a model in facealignment so we need to change code in a proj ect. The solution is to change a line 16 to self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType.TWO_D) in /content/video-retalking/third_part/face3d/extract_kp_videos.py from LandmarksType._2D . Then same in line 152 in utils\alignment_stit.py .Have a nice day everyone.

This was helpful, thank you

from video-retalking.

Ok, my China friends. The problem been bc AdrianBulat change a model in facealignment so we need to change code in a proj ect. The solution is to change a line 16 to in /content/video-retalking/third_part/face3d/extract_kp_videos.py from LandmarksType._2D . Then same in line 152 in utils\alignment_stit.py .Have a nice day everyone.self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType.TWO_D)
I now have a new error with the following message

`
PS E:\video-retalking> python inference.py --face examples/face/1.mp4 --audio examples/audio/1.wav --outfile results/1_1.mp4
[Info] Using cuda for inference.
Downloading: "https://github.com/xinntao/facexlib/releases/download/v0.1.0/detection_Resnet50_Final.pth" to C:\Users\admin.conda\envs\video_retalking\lib\site-packages\facexlib\weights\detection_Resnet50_Final.pth

100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 104M/104M [02:59<00:00, 611kB/s]
Downloading: "https://github.com/xinntao/facexlib/releases/download/v0.2.2/parsing_parsenet.pth" to C:\Users\admin.conda\envs\video_retalking\lib\site-packages\facexlib\weights\parsing_parsenet.pth

100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 81.4M/81.4M [01:29<00:00, 952kB/s]
[Step 0] Number of frames available for inference: 135
[Step 1] Landmarks Extraction in Video.
landmark Det:: 1%|▉ | 1/135 [00:05<11:37, 5.20s/it]nvrtc: error: invalid value for --gpu-architecture (-arch)

nvrtc compilation failed:

#define NAN __int_as_float(0x7fffffff)
#define POS_INFINITY __int_as_float(0x7f800000)
#define NEG_INFINITY __int_as_float(0xff800000)

template
device T maximum(T a, T b) {
}

template
device T minimum(T a, T b) {
return isnan(a) ? a : (a < b ? a : b);
}

extern "C" global
void fused_cat_cat(float* tinput0_42, float* tinput0_46, float* tout3_67, float* tinput0_60, float* tinput0_52, float* tout3_71, float* aten_cat, float* aten_cat_1) {
{
if (blockIdx.x<512 ? 1 : 0) {
__ldg(tinput0_60 + (512 * blockIdx.x + threadIdx.x) % 262144) : __ldg(tinput0_52 + (512 * blockIdx.x + threadIdx.x) % 262144 - 131072)) : __ldg(tout3_71 + (512 * blockIdx.x + threadIdx.x) % 262144 - 196608));
}
aten_cat[512 * blockIdx.x + threadIdx.x] = ((((512 * blockIdx.x + threadIdx.x) / 4096) % 256<192 ? 1 : 0) ? ((((512 * blockIdx.x + threadIdx.x) / 4096) % 256<128 ? 1 : 0) ? __ldg(tinput0_42 + (512 * blockIdx.x + threadIdx.x) % 1048576) : __ldg(tinput0_46 + (512 * blockIdx.x + threadIdx.x) % 1048576 - 524288)) : __ldg(tout3_67 + (512 * blockIdx.x + threadIdx.x) % 1048576 - 786432));
}
}

landmark Det:: 1%|▉ | 1/135 [00:05<11:55, 5.34s/it]
Traceback (most recent call last):
File "inference.py", line 342, in
main()
File "inference.py", line 79, in main
lm = kp_extractor.extract_keypoint(frames_pil, './temp/'+base_name+'_landmarks.txt')
File "E:\video-retalking\third_part\face3d\extract_kp_videos.py", line 27, in extract_keypoint
current_kp = self.extract_keypoint(image)
File "E:\video-retalking\third_part\face3d\extract_kp_videos.py", line 55, in extract_keypoint
return keypoints
UnboundLocalError: local variable 'keypoints' referenced before assignment
`
I downgraded numpy to 1.23.1 according to their method, but it still doesn't work. Do you have this problem?

from video-retalking.

No, I dont. That looks lile u have a problem with driver or cuda instalation. Try to update cuda vs cudnn. Whats your video card? Also I have installed torch through "conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch". Now its actually another issue for you than topic starter.

from video-retalking.

No, I dont. That looks lile u have a problem with driver or cuda instalation. Try to update cuda vs cudnn. Whats your video card? Also I have installed torch through "conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch". Now its actually another issue for you than topic starter.

My graphics card is NVIDIA Geforce RTX 4090 and the CUDA version is 12.1. Because torch was slow to download in China, conda downloads often failed. So I used the pip image to download. My current torch version is 1.9.1+cu111 and torchvision version is 0.10.1+cu111

from video-retalking.

您至少需要 CUDA 11.7 火炬版本。你安装的电流支持你的真棒卡/

Thank you very much! I'll try again. I think the high version of cuda is compatible with the low version of pytorch. Just beginning to contact this knowledge! Thank you very much for your explanation!

from video-retalking.

Ok, my China friends. The problem been bc AdrianBulat change a model in facealignment so we need to change code in a proj ect. The solution is to change a line 16 to self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType.TWO_D) in /content/video-retalking/third_part/face3d/extract_kp_videos.py from LandmarksType._2D . Then same in line 152 in utils\alignment_stit.py .Have a nice day everyone.

it fixes my problem. You are my hero buddy

from video-retalking.

Ok, my China friends. The problem been bc AdrianBulat change a model in facealignment so we need to change code in a proj ect. The solution is to change a line 16 to self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType.TWO_D) in /content/video-retalking/third_part/face3d/extract_kp_videos.py from LandmarksType._2D . Then same in line 152 in utils\alignment_stit.py .Have a nice day everyone.

much appreciate！！ this is the solution. thank you .

from video-retalking.