You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In your code there any many "print" messages, which are good for the development, but they can cause problem with "VSPipe" and in general with any "pipe" solution because too many messages in few time can broken the "pipe".
I experienced that with the CUDA version of "dlib" where the messages are written every 0.4 secs. It seems that the CPU version of "dlib" is not causing this kind of problem because the messages are written every 20 secs.
With the CUDA enabled version the are some clips where the encoding stop with the following error message:
To fix this issue I commented out the "print" in the following lines
->face_restoration_helper.py
Line 183: print('No face detected. Try to increase upsample_num_times.')
Line 187: print('Detect several faces and only keep the largest.') ->matlab_cp2tform.py
Line 282: print('\n--->uv:')
Line 283: print(uv)
Line 284: print('\n--->xy:')
Line 285: print(xy)
Line 289: print('\n--->trans matrix:')
Line 290: print(trans)
Line 292: print('\n--->trans_inv matrix:')
Line 293: print(trans_inv)
Line 295: print('\n---> apply transform to uv')
Line 296: print('\nxy_m = uv_augmented * trans')
Line 299: print(xy_m)
Line 301: print('\nxy_m = tformfwd(trans, uv)')
Line 303: print(xy_m)
Line 305: print('\n---> apply inverse transform to xy')
Line 306: print('\nuv_m = xy_augmented * trans_inv')
Line 309: print(uv_m)
Line 311: print('\nuv_m = tformfwd(trans_inv, xy)')
Line 313: print(uv_m)
Line 316: print('\nuv_m = tforminv(trans, xy)')
Line 317: print(uv_m)
and then cleared the cache.
Now the vspipe is working as expected.
Please apply the same changes on your version, or add an option in your filter to disable the debug messages.
Thanks
Dan
The text was updated successfully, but these errors were encountered:
In your code there any many "print" messages, which are good for the development, but they can cause problem with "VSPipe" and in general with any "pipe" solution because too many messages in few time can broken the "pipe".
I experienced that with the CUDA version of "dlib" where the messages are written every 0.4 secs. It seems that the CPU version of "dlib" is not causing this kind of problem because the messages are written every 20 secs.
With the CUDA enabled version the are some clips where the encoding stop with the following error message:
Error: fwrite() call failed when writing frame: xxx, plane: 0, errno: 32
To fix this issue I commented out the "print" in the following lines
->face_restoration_helper.py
Line 183: print('No face detected. Try to increase upsample_num_times.')
Line 187: print('Detect several faces and only keep the largest.')
->matlab_cp2tform.py
Line 282: print('\n--->uv:')
Line 283: print(uv)
Line 284: print('\n--->xy:')
Line 285: print(xy)
Line 289: print('\n--->trans matrix:')
Line 290: print(trans)
Line 292: print('\n--->trans_inv matrix:')
Line 293: print(trans_inv)
Line 295: print('\n---> apply transform to uv')
Line 296: print('\nxy_m = uv_augmented * trans')
Line 299: print(xy_m)
Line 301: print('\nxy_m = tformfwd(trans, uv)')
Line 303: print(xy_m)
Line 305: print('\n---> apply inverse transform to xy')
Line 306: print('\nuv_m = xy_augmented * trans_inv')
Line 309: print(uv_m)
Line 311: print('\nuv_m = tformfwd(trans_inv, xy)')
Line 313: print(uv_m)
Line 316: print('\nuv_m = tforminv(trans, xy)')
Line 317: print(uv_m)
and then cleared the cache.
Now the vspipe is working as expected.
Please apply the same changes on your version, or add an option in your filter to disable the debug messages.
Thanks
Dan
The text was updated successfully, but these errors were encountered: