YOLO / yolo /utils /model_utils.py

Commit History

πŸ› [Fix] momentum scaling bugs
c0e2436

henry000 commited on

πŸ”₯ [Remove] & fix typo of momentum schedule
b4dad5e

henry000 commited on

πŸ”§ [Update] EMA with sync state_dict
1d404e2

henry000 commited on

πŸš‘οΈ [Fix] broadcast of EMA and sync_dist only in val
89a6526

henry000 commited on

🚸 [Add] try-except in loading cache files
1690354

henry000 commited on

✨ [Add] new EMA! for torch lightning
c4cd90a

henry000 commited on

✨ [Add] a momentum schedule for wramup epoch
4be6676

henry000 commited on

✨ [Add] dynamic image size loader
aba5422

henry000 commited on

✏️ [Fix] Process <- Proccess typo
abc3992

henry000 commited on

🌐 [Add] logging of optimizre(learning rate)
5fcc6be

henry000 commited on

✨ [Update] back the warm up batch step
604c897

henry000 commited on

πŸ› [Update] Momentum of the optimizer
42e909b

henry000 commited on

✨ [New] use lightning framework to training!
8b3b3ef

henry000 commited on

πŸ’¬ [Update] logging -> loguru, consistency logging
0174b5b

henry000 commited on

✨ [Add] mutli-GPU on validation mode
e0c8580

henry000 commited on

πŸ› [Fix] LR schedule bugs, batchidx should increase
10420ed

henry000 commited on

πŸ”¨ [Remove] device in loss func, not sure correct
60c4943

henry000 commited on

⚑️ [Update] Optimizer, with more fine LR
5b9da41

henry000 commited on

πŸ’š [Fix] MPS CI bugs, enable using MPS
4056352

henry000 commited on

βœ… [Pass] test with fixup choosing device
ecc08c8

henry000 commited on

πŸš‘οΈ [Fix] get_device typo, not return cpu
3b31306

henry000 commited on

πŸ”¨ [Add] Anc2Box conveter, for YOLOv7's output
d5a73bd

henry000 commited on

βœ… [Pass] test with lower system requirements
2ee7407

henry000 commited on

πŸ”€ [Merge] branch 'TRAIN' into TEST
1ded329

henry000 commited on

πŸ› [Fix] #32 bugs, turn 8400->auto anchor size
7f8ebc5

henry000 commited on

✨ [New] local mAP calculation, AP.5:.95 - mAP.5
21a413f

henry000 commited on

πŸ”¨ [Add] the PostProccess class to convert predicts
475302b

henry000 commited on

πŸš€ [New] DDP mode for training model
1fe2937

henry000 commited on

🚧 [WIP] DDP for model training
7d7e199

henry000 commited on

πŸ”§ [Update] all structure of yaml, more readability
97e9dcb

henry000 commited on

🚚 [Rename] All utils, tools, higher readability
dcceddd

henry000 commited on