MobileFaceSwap
MobileFaceSwap copied to clipboard
Seanseattle
About FPS
I run it on 3090 but the FPS can't reach 25fps Am I missing some details? Looking forward to your response
I run it on 3090 but the FPS can't reach 25fps Am I missing some details? Looking forward to your response
Hi, could you tell me your FPS on 3090?
I run it on 3090 but the FPS can't reach 25fps Am I missing some details? Looking forward to your response
Hi, could you tell me your FPS on 3090?
As shown below, the processing speed of each image is between 0.15 seconds and 0.2 seconds
python time_check.py --target_video_path 10s_mobilefaceswap.mp4 --source_img_path /mnt/mfs/wangdonghui/data/1.jpg --use_gpu True
W1019 11:25:16.988435 15542 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 8.6, Driver API Version: 11.7, Runtime API Version: 11.6
W1019 11:25:17.036190 15542 gpu_resources.cc:91] device: 0, cuDNN Version: 8.2.
input mean and std: 127.5 127.5
find model: ./checkpoints/landmarks/glintr100.onnx recognition
find model: ./checkpoints/landmarks/scrfd_10g_bnkps.onnx detection
set det-size: (640, 640)
/mnt/mfs/wangdonghui/MobileFaceSwap/utils/align_face.py:30: FutureWarning: rcond parameter will change to the default of machine precision times max(M, N) where M and N are the input matrix dimensions.
To use the future default and silence this warning we advise to pass rcond=None, to keep using the old, explicitly pass rcond=-1.
A, res, rank, s = np.linalg.lstsq(src_pts_, dst_pts_)
W1019 11:26:06.712710 15542 gpu_resources.cc:201] WARNING: device: . The installed Paddle is compiled with CUDNN 8.4, but CUDNN version in your machine is 8.2, which may cause serious incompatible bug. Please recompile or reinstall Paddle with compatible CUDNN version.
0%| | 0/250 [00:00<?, ?it/s]
0.22238469123840332
0%|▍ | 1/250 [00:01<07:06, 1.71s/it]
0.18510675430297852
1%|▉ | 2/250 [00:02<03:37, 1.14it/s]
0.18127012252807617
1%|█▎ | 3/250 [00:02<02:26, 1.68it/s]
0.19201445579528809
2%|█▊ | 4/250 [00:02<02:08, 1.92it/s]
0.15575170516967773
2%|██▏ | 5/250 [00:03<02:20, 1.75it/s]
0.17618131637573242
2%|██▋ | 6/250 [00:03<01:53, 2.15it/s]
0.18538236618041992
3%|███ | 7/250 [00:03<01:37, 2.50it/s]
0.1773383617401123
3%|███▌ | 8/250 [00:04<01:40, 2.42it/s]
0.1560380458831787
4%|███▉ | 9/250 [00:05<02:35, 1.55it/s]
0.14922428131103516
4%|████▎ | 10/250 [00:05<02:04, 1.93it/s]
0.15995073318481445
4%|████▊ | 11/250 [00:06<01:57, 2.04it/s]
0.14891815185546875
5%|█████▏ | 12/250 [00:06<01:37, 2.44it/s]
0.14820241928100586
5%|█████▋ | 13/250 [00:07<02:21, 1.68it/s]
0.1681513786315918
6%|██████ | 14/250 [00:07<02:03, 1.92it/s]
0.1500542163848877
6%|██████▌ | 15/250 [00:08<01:45, 2.22it/s]
0.14726758003234863
6%|██████▉ | 16/250 [00:08<01:36, 2.42it/s]
0.14962410926818848
7%|███████▍ | 17/250 [00:08<01:48, 2.14it/s]
0.15721940994262695
7%|███████▊ | 18/250 [00:09<01:31, 2.53it/s]
0.1529240608215332
7%|███████▊ | 18/250 [00:10<02:19, 1.67it/s]
Traceback (most recent call last):
File "time_check.py", line 96, in
Is it a problem with my environment? This is a new environment, the speed can reach 10FPS in the old environment, I think this should be one of the problems affecting the inference speed
My test inference model on A100 only gets 17fps at most, am I doing something wrong?
