603ea0bfdc
update log_dir to runs/exp #107
5 years ago
72d5b58b9a
disable LR plot to suppress warning message
5 years ago
24c5a941f0
--resume EMA fix #292
5 years ago
9d631408a2
Move hyp and opt yaml save to top of train()
...
Fixes bug where scaled values were saved in hyp.yaml, which would cause continuity issues with --resume
5 years ago
bf6f41567a
hyperparameter printout update
5 years ago
dc5e18390a
Merge branch 'master' into advanced_logging
5 years ago
6b134d93c5
Update train.py
5 years ago
16f6834486
update train.py and experimental.py
5 years ago
52bac22f09
Add in --resume functionality with option to specify path or to get most recent run
5 years ago
f517ba81c7
Merge branch 'master' into advanced_logging
5 years ago
956511dafd
fix LR bug
5 years ago
bb3c346916
model.yaml nc inherited from dataset.yaml
5 years ago
df224a0d8f
EMA bug fix #279
5 years ago
3bdea3f697
strip_optimizer() bug fix #253
5 years ago
e02a189a3a
Merge pull request #245 from yxNONG/patch-2
...
Unify the check point of single and multi GPU
5 years ago
597ed4ce63
Update train.py
5 years ago
13f69777a6
typo fix
5 years ago
1aa2b67933
Update train.py
5 years ago
86784cfdbf
--resume bug fix #252
5 years ago
ad4c22cbfe
--resume bug fix #187
5 years ago
3b16c865f0
assert --epochs
5 years ago
cdb9bde181
Unify the check point of single and multi GPU
...
save the model.hyp etc to checkpoint when use multi GPU training
5 years ago
b203c9b7ff
update train.py incompatible model message fix #222
5 years ago
37e13f8846
update mosaic border
5 years ago
e18e6811dc
Merge branch 'master' into advanced_logging
5 years ago
22fb2b0c25
refactor dataloader
5 years ago
256a3e89d2
small dataset bug fix #140
5 years ago
b50fdf16af
model.names multi-GPU bug fix #94
5 years ago
de191655e4
Fix yaml saving (don't sort keys), reorder --opt keys, bug fix hyp dict accessor
5 years ago
2d396bea00
Fix bug in --help from percent sign in help string
5 years ago
b8557f87e3
add stride to datasets.py
5 years ago
611aacf1bf
Turn opt into dictionary before sending it to yaml
5 years ago
bc4ef4861b
Default optimizer SGD
5 years ago
7abf202cad
Mode all optimizer settings to 'hyp.yaml', integrate proper momentum with Adam optimizer
5 years ago
7edbf6570e
Fix help message for cfg files
5 years ago
d64ad0fbf3
Remove --resume functionality and related checks/logic.
5 years ago
6c1b87a42e
update google_utils import
5 years ago
1f1917ef56
remove fast, add merge
5 years ago
e572bb0803
Add plot_results save location to log_dir
5 years ago
899f1d4bde
Fix DDP bug in single process multiple device use cases
5 years ago
cdf1eac9f7
Merge pull request #107 from Lornatang/fix-reference-bugs
...
fix refrence bug
5 years ago
cce95e744d
backbone as FP16, save default to FP32
5 years ago
d9b64c27c2
save ckpt in FP16 #119
5 years ago
9fdb0fbacf
AutoAnchor bug fix # 117
5 years ago
c8152c81a6
Syntax fixes
5 years ago
9b7386f603
Add save_dir arg to test.test, use arg as location for saving batch jpgs
5 years ago
945307beba
Add save_dir to plot_lr_scheduler and plot_labels
...
Set save_dir = log_dir in train.py
5 years ago
3b2b330872
Move results.txt from weights/ to log_dir
5 years ago
ade023cff2
Fix hyp file read in and dict update.
...
Add example of hyp yaml
5 years ago
2368603484
fix refrence bug
...
In torch==1.5, the import of the API has changed. Although it does not interrupt the operation of the program, it seems to me to be an implicit error and may throw an exception in later versions.
5 years ago
8db51c7002
tb_writer bug fix
5 years ago
5f2eeba233
remove old print statements
5 years ago
afe1df385b
dist.destroy_process_group() bug fix
5 years ago
333f678b37
add update default hyp dict with provided yaml
5 years ago
a448c3bcd7
add logic for resuming and getting hyp for resume run
5 years ago
25e51bcec7
add util function to get most recent last.pt file
...
added logic in train.py __main__ to handle resuming from a run
5 years ago
490f1e7b9c
add save_dir arg to plot_lr_scheduler, default to current dir.
...
Uncomment plot_lr_scheduler in train() and pass log_dir as save location
5 years ago
4418809cf5
change weights dir (wdir) to be unique to each run, under log_dir
5 years ago
d9f446cd81
add save yaml of opt and hyp to tensorboard log_dir in train()
5 years ago
a85e6d0fc0
add parser arg for hyp yaml file
5 years ago
5a50491fa1
check_anchors bug fix
5 years ago
05b8ee5ca4
check_anchors() bug fix #102
5 years ago
ec81c7b5f2
check_anchors() bug fix #90
5 years ago
8b26e89006
AutoAnchor bug fix #72
5 years ago
bdd9fee841
update fast mode
5 years ago
1c0b6236e3
update fast mode
5 years ago
915b1481fc
default check_git_status() to True
5 years ago
14523bb030
FP16 to FP32 ckpt load
5 years ago
c5966abba8
glob search bug fix #77
5 years ago
31f3310029
assert best possible recall > 0.9 before training
5 years ago
c3d4d321d3
Ensures weights/ dir exists
...
Allows train.py to be run outside of yolov5/ directory.
5 years ago
099e6f5ebd
--img-size stride-multiple verification
5 years ago
22d6088205
speed-reproducibility fix #17
5 years ago
55ca5c74d2
multi-scale fix #16
5 years ago
7c2832cd49
assert equal model and dataset classes
5 years ago
11121e39ed
updates
5 years ago
eb97b2e413
NMS fast mode
5 years ago
ce36905358
updates
5 years ago
1e84a23f38
initial commit
5 years ago
Glenn Jocher
Alex Stoken
Laughing
yxNONG
Lornatang
Lauritzen Kasper Primdal