level 1
李子健
楼主
报错内容:“
This section contains the details the devs need to fix this issue.
If you are reporting a bug, please include this entire section.
If you leave out any of it, there is a good chance the devs will not be able to help.
Error: received a OSError at stage "predict".
Message: "Failed to open file b'C:\\Users\\\xe5\x88\x98\xe5\xb0\x8f\xe6\xb3\xbd\\AppData\\Local\\Temp\\scipy-hk4tvvuu'"
Full traceback:
File "afy\yanderify.py", line 134, in worker_thread
prediction = predictor.predict(frame)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 77, in predict
self.start_frame_kp = self.get_frame_kp(driving_frame)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 95, in get_frame_kp
kp_image = self.normalize_alignment_kp(kp_image)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 103, in normalize_alignment_kp
area = ConvexHull(kp[:, :2]).volume
File "qhull.pyx", line 2335, in scipy.spatial.qhull.ConvexHull.__init__
File "qhull.pyx", line 266, in scipy.spatial.qhull._Qhull.__init__
File "messagestream.pyx", line 36, in scipy._lib.messagestream.MessageStream.__init__
<log>
device=cpu
</log>
<inputs>
E:/cudachange/sucai.mp4 E:/cudachange/taogeimage.png E:/cudachange/NVIDIA.mp4
</inputs>
This is the last line of the crash report section.
===========================
yanderify crashed!
some common problems:
- you have an AMD card. AMD cards are not supported in GPU mode for technical reasons. However, you can run in CPU mode, albeit much lower. Please read the disclaimer at the top about CPUs!
- you have an NVIDIA card, but there is either not enough VRAM or the card is too old. >=700-series cards with >=2GB dedicated VRAM should work fine
- you have a working card, but there is not enough available VRAM to run the deepfake process. Browsers commonly cause VRAM issues. If you have any games open, try closing them.
- one of the devs ****ed up somewhere. if that is the case, make sure to submit the full crash report (you might have to scroll up!), otherwise we cannot help you!
”
重点我开了USE CPU没用,而且我是GTX 1650S,打开时有3.5GB显存剩余啊没救了?谢大佬,不行去(退)图(显)吧(卡
)
2020年10月07日 09点10分
1
This section contains the details the devs need to fix this issue.
If you are reporting a bug, please include this entire section.
If you leave out any of it, there is a good chance the devs will not be able to help.
Error: received a OSError at stage "predict".
Message: "Failed to open file b'C:\\Users\\\xe5\x88\x98\xe5\xb0\x8f\xe6\xb3\xbd\\AppData\\Local\\Temp\\scipy-hk4tvvuu'"
Full traceback:
File "afy\yanderify.py", line 134, in worker_thread
prediction = predictor.predict(frame)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 77, in predict
self.start_frame_kp = self.get_frame_kp(driving_frame)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 95, in get_frame_kp
kp_image = self.normalize_alignment_kp(kp_image)
File "E:\BaiduNetdiskDownload\yanderify\yanderify\yanderify\cuda100\afy\afy\predictor_local.py", line 103, in normalize_alignment_kp
area = ConvexHull(kp[:, :2]).volume
File "qhull.pyx", line 2335, in scipy.spatial.qhull.ConvexHull.__init__
File "qhull.pyx", line 266, in scipy.spatial.qhull._Qhull.__init__
File "messagestream.pyx", line 36, in scipy._lib.messagestream.MessageStream.__init__
<log>
device=cpu
</log>
<inputs>
E:/cudachange/sucai.mp4 E:/cudachange/taogeimage.png E:/cudachange/NVIDIA.mp4
</inputs>
This is the last line of the crash report section.
===========================
yanderify crashed!
some common problems:
- you have an AMD card. AMD cards are not supported in GPU mode for technical reasons. However, you can run in CPU mode, albeit much lower. Please read the disclaimer at the top about CPUs!
- you have an NVIDIA card, but there is either not enough VRAM or the card is too old. >=700-series cards with >=2GB dedicated VRAM should work fine
- you have a working card, but there is not enough available VRAM to run the deepfake process. Browsers commonly cause VRAM issues. If you have any games open, try closing them.
- one of the devs ****ed up somewhere. if that is the case, make sure to submit the full crash report (you might have to scroll up!), otherwise we cannot help you!
”
重点我开了USE CPU没用,而且我是GTX 1650S,打开时有3.5GB显存剩余啊没救了?谢大佬,不行去(退)图(显)吧(卡