Session is now saved to the model folder.
blur and erode ranges are increased to -400+400
hist-match-bw is now replaced with seamless2 mode.
Added 'ebs' color transfer mode (works only on Windows).
FANSEG model (used in FAN-x mask modes) is retrained with new model configuration
and now produces better precision and less jitter
if input frames are changed (amount or filenames)
then interactive converter automatically starts a new session.
if model is more trained then all frames will be recomputed again with their saved configs.
ConvertAvatar: fix input image after fix landmarks face align,
VideoEd: video_from_sequence now uses pipe input to input any filenames instead of %.5d. formatted
With interactive converter you can change any parameter of any frame and see the result in real time.
Converter: added motion_blur_power param.
Motion blur is applied by precomputed motion vectors.
So the moving face will look more realistic.
RecycleGAN model is removed.
Added experimental AVATAR model. Minimum required VRAM is 6GB (NVIDIA), 12GB (AMD)
Usage:
1) place data_src.mp4 10-20min square resolution video of news reporter sitting at the table with static background,
other faces should not appear in frames.
2) process "extract images from video data_src.bat" with FULL fps
3) place data_dst.mp4 video of face who will control the src face
4) process "extract images from video data_dst FULL FPS.bat"
5) process "data_src mark faces S3FD best GPU.bat"
6) process "data_dst extract unaligned faces S3FD best GPU.bat"
7) train AVATAR.bat stage 1, tune batch size to maximum for your card (32 for 6GB), train to 50k+ iters.
8) train AVATAR.bat stage 2, tune batch size to maximum for your card (4 for 6GB), train to decent sharpness.
9) convert AVATAR.bat
10) converted to mp4.bat
updated versions of modules
Fixing an issue caused by attempting to sort Path objects. Directly using `<` is unsupported between these, so `sorted()` needs a key specified.
"PurePath" objects support `>` while normal paths do not, causing the confusion.
https://docs.python.org/3/library/pathlib.html
Linux does not guarantee filenames are returned in any specific order. This leads to exporting frames in random order, sorting them here makes the export run sequentially. Other portions of the program should remain unaffected, if not behave more consistently (E.G. get_first_file_by_stem).
This mostly helpful during exporting. Say you are expecting to not have faces for frames 1000-2000, during your export all the "no faces for..." messages will appear in random order. Since you are expecting to see this you ignore them. If you are also (unexpectedly) missing a face for frame 3000 you will not head the warning since it's mixed up in all the warnings that you are expecting. With this patch export runs in sequential order, you'll see the messages all in a row for frames 1000-2000, then again at 3000. The user is much more likely to see and head the warning this way.
This also allows you to force stop the export midway though and have a contiguous set of frames to encode and preview.
An issue affecting at least 2070 and 2080 cards (possibly other RTX cards too) requires auto growth to be enabled for TensorFlow to work.
I don't know enough about the impact of this change to know whether this ought to be made optional or not, but for RTX owners, this simple change fixes TensorFlow errors when generating models.
Enable autobackup? (y/n ?:help skip:%s) :
Autobackup model files with preview every hour for last 15 hours. Latest backup located in model/<>_autobackups/01
SAE: added option only for CUDA builds:
Enable gradient clipping? (y/n, ?:help skip:%s) :
Gradient clipping reduces chance of model collapse, sacrificing speed of training.
Colombo