365 Commits (master)

Author SHA1 Message Date
  Glenn Jocher eb1217f3ac
Add PyTorch AMP check (#7917) 2 years ago
  Glenn Jocher f00071416f
Refactor collections and fstrings (#7821) 2 years ago
  Glenn Jocher 4a295b1a89
Add `@threaded` decorator (#7813) 2 years ago
  Glenn Jocher 9d8ed37df7
Rename `utils/datasets.py` > `utils/dataloaders.py` (#7799) 2 years ago
  Glenn Jocher d95a728f55
Implement DDP `static_graph=True` (#6940) 2 years ago
  Glenn Jocher aa7a0e91b1
Update `on_train_end` callback (#7716) 2 years ago
  Glenn Jocher b53917de8d
Remove `tqdm.auto` (#7599) 2 years ago
  Glenn Jocher 7043872f25
PyTorch Hub `_verbose=False` fix2 (#7550) 2 years ago
  Ayush Chaurasia 4cac1cfc8f
Add `--noplots` flag to suppress figures and images logging (#7534) 2 years ago
  Glenn Jocher 23718df1c6
Fix val `plots=plots` (#7524) 2 years ago
  Glenn Jocher c9042dc2ad
Improved non-latin `Annotator()` plotting (#7488) 2 years ago
  Glenn Jocher 4bb7eb8b84
Dynamic normalization layer selection (#7392) 2 years ago
  Vardan Agarwal fa569cdae5
Add support for different normalization layers (#7377) 2 years ago
  Glenn Jocher bd2dda8e64
Update optimizer param group strategy (#7376) 2 years ago
  Glenn Jocher 5f941a84ef
Print dataset scan only `if RANK in (-1, 0)` (#7337) 2 years ago
  Glenn Jocher 245d6459a9
Add callbacks (#7315) 2 years ago
  Glenn Jocher f735458987
Use `tqdm.auto` (#7311) 2 years ago
  Glenn Jocher 2c3221844b
CLI `fire` prep updates (#7229) 2 years ago
  Jirka Borovec c3d5ac151e
precommit: yapf (#5494) 2 years ago
  Glenn Jocher df9008ee69
Add train.py `--name cfg` option (#7202) 2 years ago
  Glenn Jocher 47f265dde2
Update min warmup iterations from 1k to 100 (#6768) 2 years ago
  Glenn Jocher 2692e67c5f
Updated VOC hyperparameters (#6732) 2 years ago
  Glenn Jocher a936f5f219
Switch default LR scheduler from cos to linear (#6729) 2 years ago
  Glenn Jocher 3670d91b05
Fix `--evolve --bucket gs://...` (#6698) 2 years ago
  Glenn Jocher 7b80545e8e
Add `--cache val` (#6663) 2 years ago
  Glenn Jocher 6769021901
Validate with 2x `--workers` (#6658) 2 years ago
  Glenn Jocher a5c9057dcc
Update train.py 2 years ago
  Christoph Fürbacher 9cf80b7f60
return `opt` from `train.run()` (#6581) 2 years ago
  Matthias Vadcard f40854b61b
Allow custom` --evolve` project names (#6567) 2 years ago
  Glenn Jocher 006eb40ab7
Improved AutoBatch DDP error message (#6568) 2 years ago
  bilzard aff0281969
Load checkpoint on CPU instead of on GPU (#6516) 2 years ago
  Ayush Chaurasia b73c62ebc5
W&B: Remember batchsize on resuming (#6512) 2 years ago
  Matthias 8efe97719c
Add `stop_training=False` flag to callbacks (#6365) 2 years ago
  Ayush Chaurasia 9bcc32a5bf
Flush callbacks when on `--evolve` (#6374) 2 years ago
  Otfot db1f83be63
Fix `train.py` parameter groups desc error (#6318) 2 years ago
  Glenn Jocher 33a67b4918
Update P2-P7 `models/hub` variants (#6230) 2 years ago
  Glenn Jocher ec4b6dd2a3
Update export format docstrings (#6151) 2 years ago
  bilzard e1dc894364
Enable AdamW optimizer (#6152) 2 years ago
  Ayush Chaurasia db6ec66a60
W&B: track batch size after autobatch (#6039) 2 years ago
  Glenn Jocher dc54ed5763
`--freeze` fix (#6044) 2 years ago
  Felix You 361705d9be
Multi-layer capable `--freeze` argument (#6019) 2 years ago
  Glenn Jocher 7c6bae0ae6
Remove NCOLS from tqdm (#5804) 2 years ago
  Ayush Chaurasia f2ca30a407
W&B: refactor W&B tables (#5737) 2 years ago
  Glenn Jocher 4ca4aec46f
Do not save hyp.yaml and opt.yaml on evolve (#5775) 2 years ago
  Glenn Jocher 47fac9ff73
Update train.py comment to 'Model attributes' (#5670) 2 years ago
  Werner Duvaud 09d170381c
Default DataLoader `shuffle=True` for training (#5623) 2 years ago
  Glenn Jocher 7473f0f95d
DDP `WORLD_SIZE`-safe dataloader workers (#5631) 2 years ago
  Glenn Jocher 61c50199a2
Update train, val `tqdm` to fixed width (#5367) 2 years ago
  Glenn Jocher 79bca2bf64
`LOGGER` consolidation (#5569) 2 years ago
  Glenn Jocher e189fa15ea
`intersect_dicts()` in hubconf.py fix (#5542) 2 years ago