Issues
- 2
SAM doesn't seem to be doing well
#87 opened by AlexLiu14 - 0
Will Layernorm or Groupnorm cause problems?
#92 opened by archerli1 - 1
Using the step function with closure
#90 opened by mathuryash5 - 5
- 4
- 5
i found it hard to implement this optimizer on yolov5.looking forward to s.b. could do me a FAVOR. THX
#77 opened by EveningLin - 1
bypass_bn is missing
#89 opened by ghazalehtrb96 - 1
`model.no_sync()` should include the forward pass
#88 opened by yzlnew - 1
Readme.MD Usage typo issue
#86 opened by fealty94 - 1
bayesian-sam
#85 opened by keishinkickback - 3
RuntimeError: Trying to backward through the graph a second time, but the saved intermediate results have already been freed. Specify retain_graph=True when calling backward the first time.
#45 opened by XieZixiUSTC - 2
Has anyone reproduce the ViT on ImageNet results using this torch implementation?
#83 opened by yzlnew - 1
SAM yolov5
#82 opened by 1556900941lizerui - 1
Wrong Adaptive mode?
#81 opened by pdradx - 1
- 2
- 1
Using SAM with torch.cuda.amp.GradScaler
#79 opened by yossigandelsman - 3
Training Tips for multiple GPUs may be invalid!
#78 opened by crj1998 - 3
- 1
Compatibility with FSDP
#67 opened by hxu38691 - 1
Model's performance
#68 opened by superbunny38 - 1
- 1
sam install
#74 opened by dududukekeke - 3
- 1
"TypeError: __init__() missing 1 required positional argument: 'base_optimizer'" with 'ddp_sharded''
#70 opened by anicolson - 1
- 4
Gradient Clipping with SAM optimizer
#64 opened by stardust-88 - 3
Multi-network, multi-loss function using SAM
#63 opened by zhanglina94 - 3
Gradient accumulation
#62 opened by borgr - 3
- 2
May I know how to install the package?
#60 opened by Carlisle-Liu - 1
rho argument type error
#58 opened by okbalefthanded - 1
Using SAM and ASAM for RNNs can cause nan outputs
#57 opened by DaStapo - 1
- 1
AMP support
#55 opened by JeanKaddour - 5
Experiments on object detection task
#41 opened by shuo-ouyang - 3
how to train my own data?
#53 opened by Chenhait - 1
where is and how to use the ASAM?
#54 opened by Geek-lixiang - 1
Issue on weight decay
#52 opened by liujingcs - 1
key error:“old_p",when I use optim.adamw
#51 opened by daixiangzi - 2
Variable `args.label_smoothing` not being passed to `smooth_cross_entropy()
#50 opened by glennmatlin - 1
Import error on windows machine
#48 opened by rschiewer - 1
no bypass_bn.py in utility.bypass_bn
#44 opened by Elsie0908 - 2
How do we give credit?
#43 opened by ClashLuke - 2
pytorch lighting seems can't use?
#36 opened by langdaoliu - 6
Experiments on ImageNet with R18 or R50?
#40 opened by jiequancui - 3
Multi GPU
#38 opened by lthilnklover - 1
How to resume SAM in pytorch?
#34 opened by whynique - 2
- 2
Question: Why step_lr?
#35 opened by JeanKaddour