Commit History
FP16 test loss bug fix
0b514da
test during training default to FP16
a1748a8
update google_utils.py
bd3e389
FP16 to FP32 ckpt load
76ca367
FP16 to FP32 ckpt load
6fb5ff0
FP16 to FP32 ckpt load
14523bb
glob search bug fix #77
c5966ab
Merge pull request #70 from Lornatang/fix-reference-bugs
ace56eb
unverified
fix problem
54795a4
Lornatang
commited on
Merge remote-tracking branch 'origin/master'
45a17ea
half() fix #66
e7c24e0
Merge pull request #65 from Ashing00/master
6570336
unverified
add "__init__.py" file in <models> folder
62eda5e
ashing
commited on
switch default inference to FP16 on GPU
1119949
FP16 inference fix
157aff2
Merge remote-tracking branch 'origin/master'
c6fc424
FP16 inference update
260b172
Created using Colaboratory
b545ee3
pyyaml >= 5.3
629d370
Merge pull request #59 from Lijun-Yu/master
4b0e374
unverified
Update hubconf.py
4f6c0cf
unverified
path fix for torch hub
0dec8ff
pycocotools git install fix
4dd75b3
Merge pull request #56 from channeladam/master
9bf35e7
unverified
Update datasets.py
335c607
unverified
Update requirements.txt
1374581
unverified
Include yolov5x extra large model weights
f47ca20
Adam
commited on
Add more logging for diagnostics
34bf936
Adam
commited on
Fix pip requirements issues
faa6aea
Adam
commited on
best possible recall fix
93bbf1b
assert best possible recall > 0.9 before training
31f3310
Merge pull request #51 from KPLauritzen/b-ensure-weights-dir-when-training
19e68e8
unverified
Ensures weights/ dir exists
c3d4d32
Lauritzen Kasper Primdal
commited on