James Betker ee8ceed6da rework tfd13 further ... - use a gated activation layer for both attention & convs - add a relativistic learned position bias. I believe this is similar to the T5 position encodings but it is simpler and learned - get rid of prepending to the attention matrix - this doesn't really work that well. the model eventually learns to attend one of its heads to these blocks but why not just concat if it is doing that?		2022-07-20 23:28:29 -06:00
..
.idea
data	Support aac datatypes	2022-07-01 00:44:20 -06:00
models	rework tfd13 further	2022-07-20 23:28:29 -06:00
scripts	add assertions to mel generator script	2022-07-19 11:23:54 -06:00
trainer	rework tfd13 further	2022-07-20 23:28:29 -06:00
utils	misc	2022-07-08 00:37:53 -06:00
multi_modal_train.py
process_video.py	misc	2021-01-23 13:45:17 -07:00
requirements.txt	mup work checkin	2022-06-09 21:15:09 -06:00
sweep.py	misc updates	2022-05-19 13:39:32 -06:00
test.py	Add FID evaluator for diffusion models	2021-06-14 09:14:30 -06:00
train.py	misc	2022-07-20 10:19:02 -06:00
use_discriminator_as_filter.py	Various mods to support better jpeg image filtering	2021-06-25 13:16:15 -06:00