Ross Wightman e861b74cf8 Pass through --model-kwargs (and --opt-kwargs for train) from command line through to model __init__. Update some models to improve arg overlay. Cleanup along the way.		2023-01-06 16:12:33 -08:00
..
__init__.py	Pass through --model-kwargs (and --opt-kwargs for train) from command line through to model __init__. Update some models to improve arg overlay. Cleanup along the way.	2023-01-06 16:12:33 -08:00
agc.py	Initial AGC impl. Still testing.	2021-02-15 23:22:44 -08:00
checkpoint_saver.py	…
clip_grad.py	Initial AGC impl. Still testing.	2021-02-15 23:22:44 -08:00
cuda.py	Initial AGC impl. Still testing.	2021-02-15 23:22:44 -08:00
decay_batch.py	Add bulk_runner script and updates to benchmark.py and validate.py for better error handling in bulk runs (used for benchmark and validation result runs). Improved batch size decay stepping on retry...	2022-07-18 18:04:54 -07:00
distributed.py	Refactor device handling in scripts, distributed init to be less 'cuda' centric. More device args passed through where needed.	2022-09-23 16:08:59 -07:00
jit.py	disable nvfuser for jit te/legacy modes (for PT 1.12+)	2022-07-13 10:34:34 -07:00
log.py	…
metrics.py	Tweak accuracy topk safety. Fix #807	2021-08-19 14:18:53 -07:00
misc.py	Pass through --model-kwargs (and --opt-kwargs for train) from command line through to model __init__. Update some models to improve arg overlay. Cleanup along the way.	2023-01-06 16:12:33 -08:00
model.py	Fix some formatting in utils/model.py	2021-10-23 20:35:36 -07:00
model_ema.py	…
random.py	Remove commented code, add more consistent seed fn	2021-04-12 09:51:36 -07:00
summary.py	Scheduler update, add v2 factory method, support scheduling on updates instead of just epochs. Add LR to summary csv. Add lr_base scaling calculations to train script. Fix #1168	2022-10-07 10:43:04 -07:00