Commit History
Refactor optimizer initialization (#8607)
51fb467
unverified
Explicitly set `weight_decay` value (#8592)
7204c1c
unverified
Training reproducibility improvements (#8213)
27d831b
unverified
Improve `--local_rank` arg comment (#8409)
e50dc38
unverified
Implementation of Early Stopping for DDP training (#8345)
6935a54
unverified
Fix bias warmup LR init (#8356)
8ebf569
unverified
Simplify and fix `--save-period` epoch 0 (#8042)
dcf8073
unverified
Add PyTorch AMP check (#7917)
eb1217f
unverified
Refactor collections and fstrings (#7821)
f000714
unverified
Add `@threaded` decorator (#7813)
4a295b1
unverified
Rename `utils/datasets.py` > `utils/dataloaders.py` (#7799)
9d8ed37
unverified
Implement DDP `static_graph=True` (#6940)
d95a728
unverified
Update `on_train_end` callback (#7716)
aa7a0e9
unverified
Remove `tqdm.auto` (#7599)
b53917d
unverified
PyTorch Hub `_verbose=False` fix2 (#7550)
7043872
unverified
Add `--noplots` flag to suppress figures and images logging (#7534)
4cac1cf
unverified
Fix val `plots=plots` (#7524)
23718df
unverified
Improved non-latin `Annotator()` plotting (#7488)
c9042dc
unverified
Dynamic normalization layer selection (#7392)
4bb7eb8
unverified
Add support for different normalization layers (#7377)
fa569cd
unverified
Update optimizer param group strategy (#7376)
bd2dda8
unverified
Print dataset scan only `if RANK in (-1, 0)` (#7337)
5f941a8
unverified
Add callbacks (#7315)
245d645
unverified
Use `tqdm.auto` (#7311)
f735458
unverified
CLI `fire` prep updates (#7229)
2c32218
unverified
precommit: yapf (#5494)
c3d5ac1
unverified
Add train.py `--name cfg` option (#7202)
df9008e
unverified
Update min warmup iterations from 1k to 100 (#6768)
47f265d
unverified
Updated VOC hyperparameters (#6732)
2692e67
unverified
Switch default LR scheduler from cos to linear (#6729)
a936f5f
unverified
Fix `--evolve --bucket gs://...` (#6698)
3670d91
unverified
Add `--cache val` (#6663)
7b80545
unverified
Validate with 2x `--workers` (#6658)
6769021
unverified
Update train.py
a5c9057
unverified
return `opt` from `train.run()` (#6581)
9cf80b7
unverified
Christoph Fürbacher
commited on
Allow custom` --evolve` project names (#6567)
f40854b
unverified
Improved AutoBatch DDP error message (#6568)
006eb40
unverified
Load checkpoint on CPU instead of on GPU (#6516)
aff0281
unverified
W&B: Remember batchsize on resuming (#6512)
b73c62e
unverified
Add `stop_training=False` flag to callbacks (#6365)
8efe977
unverified
Flush callbacks when on `--evolve` (#6374)
9bcc32a
unverified
Fix `train.py` parameter groups desc error (#6318)
db1f83b
unverified
Update P2-P7 `models/hub` variants (#6230)
33a67b4
unverified
Update export format docstrings (#6151)
ec4b6dd
unverified
Enable AdamW optimizer (#6152)
e1dc894
unverified
bilzard
commited on