DL-Art-School/codes
James Betker 5f2c722a10 SRG2 revival
Big update to SRG2 architecture to pull in a lot of things that have been learned:
- Use group norm instead of batch norm
- Initialize the weights on the transformations low like is done in RRDB rather than using the scalar. Models live or die by their early stages, and this ones early stage is pretty weak
- Transform multiplexer to use u-net like architecture.
- Just use one set of configuration variables instead of a list - flat networks performed fine in this regard.
2020-07-09 17:34:51 -06:00
..
.idea
data Add feature to dataset which allows it to force images to be a certain size. 2020-07-03 15:19:16 -06:00
data_scripts Bug fixes and new gan mechanism 2020-07-08 17:40:26 -06:00
metrics
models SRG2 revival 2020-07-09 17:34:51 -06:00
options
scripts
temp
utils SRG2 revival 2020-07-09 17:34:51 -06:00
distill_torchscript.py Misc changes 2020-07-01 11:28:23 -06:00
onnx_inference.py Misc 2020-06-14 11:03:02 -06:00
process_video.py Debug process_video 2020-05-29 20:44:50 -06:00
recover_tensorboard_log.py Misc 2020-07-06 20:44:07 -06:00
requirements.txt Introduce (untested) colab mode 2020-06-01 15:09:52 -06:00
run_scripts.sh
test.py Misc 2020-06-02 09:35:52 -06:00
train.py SRG2 revival 2020-07-09 17:34:51 -06:00