facebookresearch/fvcore

Issues

• iopath dependency

  #141 opened a year ago by BagelOrb
  2

• Checkpointer load should specify `weights_only`

  #154 opened 3 months ago by baldassarreFe
  0

• FlopCountAnalysis issues

  #153 opened 4 months ago by cavalleria
  0

• different bn_stats on different DDP ranks after update_bn_stats as precise_bn does not synchronize batch_size

  #152 opened 4 months ago by guzy0324
  0

• RuntimeError: Detected that you are using FX to torch.jit.trace a dynamo-optimized function. This is not supported at the moment.

  #150 opened 8 months ago by arghavan-kpm
  0

• The flopcounter does not support LeakyReLU?

  #105 opened 3 years ago by Windaway
  2

• ValueError: You have to specify either decoder_input_ids or decoder_inputs_embeds

  #149 opened 10 months ago by Aafiya-H
  0

• ValueError: Invalid type <class 'numpy.int32'> for the flop count! Please use a wider type to avoid overflow.

  #104 opened 3 years ago by WYHZQ
  5

• Conflict with torch.distributed?

  #145 opened a year ago by Wuzimeng
  1

• flop count analysis of LSTM layers

  #98 opened 3 years ago by HendrikKlug-synthara
  2

• Unsupported operator aten::max_pool2d encountered 5 time(s)

  #148 opened a year ago by Rakeshlal791
  0

• Counting FLOPS for a custom op with set_op_handle: a toy example that doesn't work.

  #147 opened a year ago by guynich
  0

• Unsupported operators

  #143 opened a year ago by qsimeon
  0

• support div mul add

  #142 opened a year ago by yokosyun
  0

• AttributeError: module 'fvcore' has no attribute 'nn'

  #95 opened 3 years ago by yongjie-lv
  2

• Quantized model size

  #139 opened 2 years ago by prinshul
  0

• Unsupported operator aten::add_

  #138 opened 2 years ago by peterzheng98
  0

• NotImplementedError during FlopCountAnalysis if submodule decorated by torch.compile

  #137 opened 2 years ago by flishwang
  0

• Sort state_dict.keys() in _strip_prefix_if_present?

  #136 opened 2 years ago by isdanni
  0

• RuntimeError in ParamScheduler

  #81 opened 4 years ago by krishnamoorthybabu
  9

• Unsupport the high dimension inputs that larger than 5.

  #135 opened 2 years ago by hailong23-jin
  0

• Error while calculating flops of transformer model.

  #129 opened 2 years ago by NeelKanwal
  9

• Clarification in the nomenclature of FLOPs / MACs

  #133 opened 2 years ago by Vinayaka2k
  1

• Feed two batches and then dividing the ended FLOPs by 2 to avoid the batch_norm error?

  #132 opened 2 years ago by cloude-z
  0

• This error occurred when I calculated flops. What is the reason?

  #125 opened 2 years ago by fxooooooooo
  1

• Flop counter for matmul does not support matrix-vector product.

  #130 opened 2 years ago by AllenYolk
  0

• Why is the sigmoid function applied to logits before calling F.binary_cross_entropy_with_logits?

  #126 opened 2 years ago by kirill-menke
  0

• Formula for batch normalization

  #124 opened 2 years ago by unipdjahanifar
  0

• DeformConv2d not supported in FLOPs counting?

  #121 opened 2 years ago by kenmbkr
  2

• Does a single multiply op count one flop?

  #119 opened 2 years ago by jixiege
  2

• Memory leak?

  #115 opened 3 years ago by josecannete
  1

• baddbmm not supported

  #117 opened 2 years ago by rwightman
  0

• Meaning of Class <ActivationCountAnalysis>

  #112 opened 3 years ago by go-ahead-maker
  0

• Clipping only applied to image arrays that arrive in uint8 type

  #110 opened 3 years ago by Trolomoll
  0

• Wrong FLOPs count if some submodules of the model are wrapped by DataParallel

  #109 opened 3 years ago by luowyang
  0

• ShapelyDeprecationWarning in CropTransform

  #103 opened 3 years ago by bowenc0221
  0

• Matmul Flops

  #108 opened 3 years ago by breuera
  2

• Counting bias term

  #102 opened 3 years ago by noamgot
  2

• Wrong ConvTranspose flops

  #78 opened 3 years ago by manipopopo
  0

• Why not return the variable `save_file` in Line 184 directly?

  #100 opened 3 years ago by DapengFeng
  0

• Add device parameter to update_bn_stats() in precise_bn.py

  #97 opened 3 years ago by hankyul2
  0

• Getting:- ImportError: cannot import name 'FakeQuantizeBase' from 'torch.quantization'

  #89 opened 3 years ago by patrickn699
  5

• The number of FLOPs turned out to be a negative number

  #93 opened 3 years ago by Asagami-Fujino
  5

• New import breaks compatibility with python 3.6/7

  #94 opened 3 years ago by radao
  0

• How can I support custom operators?

  #83 opened 3 years ago by xunleiwu
  4

• BUG: ImportError: cannot import name 'FakeQuantizeBase' from 'torch.quantization'

  #87 opened 3 years ago by hasansalimkanmaz
  9

• can't calculate the flops of loop in def forward

  #92 opened 3 years ago by wwHwo
  3

• "flop_count" for Detectron2 DeformConv

  #84 opened 4 years ago by bowenc0221
  3

• How do I cite this repo?

  #82 opened 4 years ago by ekosman
  2

• Overall Flops are less than the modules

  #76 opened 4 years ago by DuanHuiyu
  5