Giters
Lyken17
/
pytorch-OpCounter
Count the MACs / FLOPs of your PyTorch model.
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
4802
Watchers:
30
Issues:
169
Forks:
523
Lyken17/pytorch-OpCounter Issues
count_normalization is only correct for batch_norm. wrong flops count for layernorm
Updated
a month ago
Comments count
1
question about BN
Updated
a month ago
I got different results using thop and torchinfo
Updated
2 months ago
Comments count
1
thop/profile.py:12: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead. `if LooseVersion(torch.__version__) < LooseVersion("1.0.0"):`
Updated
2 months ago
Comments count
2
Source distribution (sdist) and Git tags
Updated
2 months ago
got 0 ops for nn.MultiheadAttention
Updated
4 months ago
Comments count
7
thop calculates torch.nn module params incorrectly
Updated
4 months ago
Comments count
1
Res add will be included when evaluate Resnet's OPs ?
Updated
4 months ago
Comments count
1
Is the latest version calculate MACs or FLOPs
Updated
7 months ago
Comments count
2
Problem in bert
Updated
7 months ago
Comments count
1
How to exclude flops of 1st input?
Updated
7 months ago
Comments count
1
RuntimeError: Can't add a new parameter
Updated
7 months ago
rename calculate_conv2d_flops
Updated
8 months ago
Comments count
1
Is thop also effective for calculating Flops for spiking neural networks?
Updated
9 months ago
will torch.matmul regards as zero_ops ?
Updated
a year ago
Incorrect macs without specifying batch size for conv layers
Updated
a year ago
How to calculate the FLOPs of each type of layers?
Updated
a year ago
How do I calculate the FLOPs of a model with some frozen layers during training?
Updated
a year ago
Comments count
1
Does MACs and FLOPs count correctly for and INT8 quantized model?
Updated
a year ago
Comments count
1
multiple inputs
Updated
a year ago
Comments count
1
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!
Closed
a year ago
Upload sdist to PyPI
Updated
a year ago
Comments count
1
why these are zero ops ? relu/dropout/pool
Closed
a year ago
Count flops by a range
Updated
a year ago
[discuss] How about set profile's verbose's default value to False?
Closed
2 years ago
[feature] Can it support only calculate the parameters which `requires_grad == True`?
Closed
2 years ago
Comments count
2
question about the MACs of nn.BatchNorm2d
Updated
2 years ago
Does it support non-image models?
Updated
2 years ago
Comments count
1
AttributeError: 'PReLU' object has no attribute 'total_params'
Updated
2 years ago
Is MACs equal to Flops?
Updated
2 years ago
Comments count
2
A suggestion about README.
Closed
2 years ago
AttributeError: 'LayerNorm' object has no attribute 'affine'
Updated
2 years ago
Comments count
12
Can these newly added keys be removed after the function profile is executed?
Updated
2 years ago
Comments count
2
A typo in evaluate_famous_models.py
Closed
2 years ago
Comments count
2
Does this library work with einops?
Closed
2 years ago
Comments count
2
F.conv2d
Closed
2 years ago
Comments count
3
AttributeError: module 'thop' has no attribute '__version__'
Closed
2 years ago
Comments count
2
Does it support torch.einsum for FLOPs?
Closed
2 years ago
Comments count
1
Does it think about to support torch.nn.PixelShuffle? It is widely used in Super-Resolution.
Closed
2 years ago
Comments count
4
/opt/conda/lib/python3.8/site-packages/thop/vision/basic_hooks.py:115: UserWarning: __floordiv__ is deprecated, and its behavior will change in a future version of pytorch.
Closed
2 years ago
Comments count
3
how to define the rule for 3rd party module?
Updated
2 years ago
Comments count
1
Are MACs values equal across devices (GPU/CPU)?
Closed
2 years ago
Comments count
1
Not compatible with python3.10
Updated
2 years ago
Comments count
7
When I calculate the flops of Torch.nn.upsample, I found that Hook in the Forward process will override the output of UpSample. Is this BUG or a special design?
Closed
2 years ago
Comments count
4
[BUG Report] thop not support AdaptiveAvgPool2d when None in the shape parameter
Updated
2 years ago
Comments count
1
Is matmul operation automatically considered?
Closed
2 years ago
Comments count
3
Mish activation function and SiLU are not supported
Updated
2 years ago
Different results when reusing the defined nn.LeakyReLU()
Updated
2 years ago
Comments count
1
FLOPs/MACs calculation of complicated operations
Closed
2 years ago
Comments count
2
Help me to check my custom op calculation code
Closed
2 years ago
Comments count
5
Previous
Next