vrtc: error: invalid value for --gpu-architecture (-arch)

[giandre]

I am getting stuck in the following error. I am trying to figure out what to do:

(video_retalking) H:\z-aitools\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.
[Step 0] Number of frames available for inference: 135
[Step 1] Landmarks Extraction in Video.
landmark Det:: 1%|▍ | 1/135 [00:14<31:21, 14.04s/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) {
return isnan(a) ? a : (a > b ? a : 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) {
aten_cat_1[512 * blockIdx.x + threadIdx.x] = ((((512 * blockIdx.x + threadIdx.x) / 1024) % 256<192 ? 1 : 0) ? ((((512 * blockIdx.x + threadIdx.x) / 1024) % 256<128 ? 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 [00:14<31:57, 14.31s/it]
Traceback (most recent call last):
File "inference.py", line 347, in
main()
File "inference.py", line 84, in main
lm = kp_extractor.extract_keypoint(frames_pil, './temp/'+base_name+'_landmarks.txt')
File "H:\z-aitools\video-retalking\third_part\face3d\extract_kp_videos.py", line 29, in extract_keypoint
current_kp = self.extract_keypoint(image)
File "H:\z-aitools\video-retalking\third_part\face3d\extract_kp_videos.py", line 57, in extract_keypoint
return keypoints
UnboundLocalError: local variable 'keypoints' referenced before assignment

I did some troubleshooting with ChatGPT and this is my current CUDA settings since I am suspecting is related to that? I am using a RTX4090.
(video_retalking) H:\z-aitools\video-retalking>python test.py
Is CUDA available: True
CUDA device count: 1
CUDA current device: 0
CUDA device name: NVIDIA GeForce RTX 4090
CUDA arch list: ['sm_37', 'sm_50', 'sm_60', 'sm_61', 'sm_70', 'sm_75', 'sm_80', 'sm_86', 'compute_37']

I appreciate any help!

[lordlxh]

renew pytorch and cuda maybe helpful.

[giandre]

I was able to fix it. I do not recall how but this process takes way too long to complete so I decided to continue trying with Dinet and Sadtalker..