vacancy Synchronized-BatchNorm-PyTorch issues

vacancy / Synchronized-BatchNorm-PyTorch

Synchronized Batch Normalization implementation in PyTorch.

MIT License

1.5k stars 189 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

spam

#50 QXPOMBA closed 1 year ago
0
.

#49 LSKhappychild closed 2 years ago
0
how to export with onnx

#48 tongchangD opened 2 years ago
0
Where is "track_running_stats" implementation code?

#47 junha1125 closed 3 years ago
1
Training stuck with multiple call of forward function

#46 ArmastusChen opened 3 years ago
7
raining couldn 't start

#45 zwyking closed 3 years ago
2
Train Stucked

#44 YinengXiong opened 3 years ago
4
a question about the highlight "use sqrt(max(var, eps)) instead of sqrt(var + eps)"

#43 louyj136 closed 3 years ago
4
batchnor while using distributed dataparallel

#42 ParthaEth closed 3 years ago
1
Training cannot start

#41 shuuchen closed 4 years ago
0
Training cannot start

#40 shuuchen opened 4 years ago
7
module问题

#39 fanjing136 closed 3 years ago
0
Question on `sqrt(max(var, eps))`

#38 nmhkahn closed 4 years ago
2
Wired things, module must have its parameters and buffers on device cuda:0 (device_ids[0]) but found one of them on device: cpu

#37 kehuantiantang closed 4 years ago
1
compatible with pytorch 1.4.0

#36 wang93 closed 4 years ago
1
Thinking about 'sync_batchnorm.batchnorm.convert_model(module)'..

#35 chrisway613 closed 4 years ago
1
RuntimeError with convert_model - "found one of them on device: cpu"

#34 FreemanG closed 4 years ago
2
How to use it when testing

#33 zhulanyun closed 4 years ago
1
How to cite this repo in bib?

#32 Anikily opened 5 years ago
0
Is this a bug that channel between input tensor and sync batchnorm are mismatch the code still run successful?

#31 lscelory closed 3 years ago
2
test gap between training and test

#30 ZhiweiYan-96 opened 5 years ago
3
Where is the "BatchNormReimpl"?

#29 skanderbug closed 5 years ago
1
Network performance is getting worse

#28 charlesCXK opened 5 years ago
2
Fix a error in docs of convert_model

#27 DTennant closed 4 years ago
1
about convert_model

#26 Re3write closed 5 years ago
6
add function convert_model to convert a model for easy usage

#25 DTennant closed 5 years ago
3
is everyone test the code in pytorch 1.0 ?

#24 wadesunyang closed 3 years ago
10
Different from torch.nn.DataParallel that will separate total batch to each gpu equally?

#23 ksnzh closed 5 years ago
1
about fp16

#22 666zz666 opened 5 years ago
5
About python test scripts and SyncBN usage

#21 Spritea closed 5 years ago
1
Can't test successfully using the scripts in ./tests

#20 Yochengliu closed 5 years ago
2
shall I use your `DataParallelWithCallback` instead of pytorch `nn.DataParallel`?

#19 CoinCheung closed 5 years ago
1
Is it more difficult to train model?

#18 yuanfuqiang456 closed 3 years ago
4
Can one past the test ?

#17 chenchr closed 5 years ago
1
Add new numeric tests

#16 vacancy closed 6 years ago
0
New numeric test

#15 acgtyrant closed 6 years ago
2
Another test script for numerical stability

#14 acgtyrant closed 6 years ago
7
`np.fmax(npa, 1e-5))` does not consider that npa is negative

#13 acgtyrant closed 6 years ago
2
How to solve the numeric error?

#12 acgtyrant closed 6 years ago
1
When I use the sync bn on 8x GPU, I will stop sometimes.

#11 yu-changqian closed 6 years ago
2
After sync batch norm is applied, more gpu memory is consumed

#10 shachoi opened 6 years ago
9
Fix pickle error

#9 aywi closed 6 years ago
1
Update test_numeric_batchnorm.py

#8 d-li14 closed 6 years ago
0
How to load bias weights

#7 guantinglin closed 6 years ago
1
How to use this module to load the pre-trained model offered by the official PyTorch?

#6 yu-changqian closed 6 years ago
3
Unstable performance between adjacent epochs during test phase

#5 7color94 closed 6 years ago
6
The mean is not averaged realy.

#4 acgtyrant closed 6 years ago
1
Wired things happen when applied to FPN

#3 ShuLiu1993 opened 6 years ago
9
_SynchronizedBatchNorm's backward is redefined by its forward, right?

#2 acgtyrant closed 6 years ago
1
Does this support torch.nn.parallel.DistributedDataParallel?

#1 acgtyrant opened 6 years ago
18