torch-summary 1.1.7

Model summary in PyTorch, based off of the original torchsummary.

torch-summary

Keras has a neat API to view the visualization of the model which is very helpful while debugging your network. In this project, we attempt to do the same in PyTorch. The goal is to provide information complementary to what is provided by print(your_model) in PyTorch.

This is a rewritten version of the original torchsummary and torchsummaryX projects by @sksq96 and @nmhkahn. There are quite a few pull requests on the original project (which hasn't been updated in over a year), so I decided to take a stab at improving and consolidating some of the features.

This version now supports:

• RNNs, LSTMs, and other recursive layers
• Branching output to explore model layers using specified depths
• Returns ModelStatistics object to access summary data
• Configurable columns of returned data

Other features:

• Verbose mode to show specific weights and bias layers
• Accepts either input data or simply the input shape to work!
• Customizable widths and custom batch dimension.
• More comprehensive testing using pytest

Usage

pip install torch-summary

or

git clone https://github.com/tyleryep/torch-summary.git

from torchsummary import summary
summary(your_model, input_data=(C, H, W))

Documentation

"""
Summarize the given PyTorch model. Summarized information includes:
    1) output shape,
    2) kernel shape,
    3) number of the parameters
    4) operations (Mult-Adds)
Args:
    model (Module): Model to summarize
    input_data (Sequence of Sizes or Tensors):
        Example input tensor of the model (dtypes inferred from model input).
        NOTE: multiple parameters to the model should be passed as *args, NOT as a List.
        - OR -
        Shape of input data as a List/Tuple/torch.Size (dtypes must match model input,
        default is FloatTensors). NOTE: For scalar parameters, use torch.Size([]).
    branching (bool): Whether to use the branching layout for the printed output.
    depth (int): number of nested layers to traverse (e.g. Sequentials)
    verbose (int):
        0 (quiet): No output
        1 (default): Print model summary
        2 (verbose): Show weight and bias layers in full detail
    col_names (List): specify which columns to show in the output. Currently supported:
        ['output_size', 'num_params', 'kernel_size', 'mult_adds']
    col_width (int): width of each column
    dtypes (List or None): for multiple inputs or args, must specify the size of both inputs.
        You must also specify the types of each parameter here.
    batch_dim (int): batch_dimension of input data
    device (torch.Device): If specified, uses this torch device for the model and model's input.
        Else defaults to torch.cuda.is_available().
    args, kwargs: Other arguments used in `model.forward` function.
"""

Examples

Get Model Summary as String

from torchsummary import summary

model_stats = summary(your_model, input_data=(C, H, W), verbose=0)
summary_str = str(model_stats)

CNN for MNIST

class CNN(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
        self.conv2_drop = nn.Dropout2d(0.3)
        self.fc1 = nn.Linear(320, 50)
        self.fc2 = nn.Linear(50, 10)

    def forward(self, x):
        x = F.relu(F.max_pool2d(self.conv1(x), 2))
        x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
        x = x.view(-1, 320)
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        return F.log_softmax(x, dim=1)


model = CNN()
summary(model, (1, 28, 28))

------------------------------------------------------------------------------------------
Layer (type:depth-idx)                   Output Shape              Param #
==========================================================================================
├─Conv2d: 1-1                            [-1, 10, 24, 24]          260
├─Conv2d: 1-2                            [-1, 20, 8, 8]            5,020
├─Dropout2d: 1-3                         [-1, 20, 8, 8]            --
├─Linear: 1-4                            [-1, 50]                  16,050
├─Linear: 1-5                            [-1, 10]                  510
==========================================================================================
Total params: 21,840
Trainable params: 21,840
Non-trainable params: 0
------------------------------------------------------------------------------------------
Input size (MB): 0.00
Forward/backward pass size (MB): 0.05
Params size (MB): 0.08
Estimated Total Size (MB): 0.14
------------------------------------------------------------------------------------------

Multiple Inputs w/ Different Data Types

class MultipleInputNetDifferentDtypes(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc1a = nn.Linear(300, 50)
        self.fc1b = nn.Linear(50, 10)

        self.fc2a = nn.Linear(300, 50)
        self.fc2b = nn.Linear(50, 10)

    def forward(self, x1, x2):
        x1 = F.relu(self.fc1a(x1))
        x1 = self.fc1b(x1)
        x2 = x2.type(torch.float)
        x2 = F.relu(self.fc2a(x2))
        x2 = self.fc2b(x2)
        x = torch.cat((x1, x2), 0)
        return F.log_softmax(x, dim=1)


summary(model, [(1, 300), (1, 300)], dtypes=[torch.float, torch.long])

Alternatively, you can also pass in the input_data itself, and torchsummary will automatically infer the data types.

input_data = torch.randn(1, 300)
other_input_data = torch.randn(1, 300).long()
model = MultipleInputNetDifferentDtypes()

summary(model, input_data, other_input_data, ...)

Explore Different Configurations

class LSTMNet(nn.Module):
    def __init__(self, vocab_size=20, embed_dim=300, hidden_dim=512, num_layers=2):
        super().__init__()
        self.hidden_dim = hidden_dim
        self.embedding = nn.Embedding(vocab_size, embed_dim)
        self.encoder = nn.LSTM(embed_dim, hidden_dim, num_layers=num_layers)
        self.decoder = nn.Linear(hidden_dim, vocab_size)

    def forward(self, x):
        embed = self.embedding(x)
        out, hidden = self.encoder(embed)
        out = self.decoder(out)
        out = out.view(-1, out.size(2))
        return out, hidden

summary(
    LSTMNet(),
    (100,),
    dtypes=[torch.long],
    branching=False,
    verbose=2,
    col_width=16,
    col_names=["kernel_size", "output_size", "num_params", "mult_adds"],
)

--------------------------------------------------------------------------------------------------------
Layer (type:depth-idx)         Kernel Shape         Output Shape         Param #          Mult-Adds
========================================================================================================
Embedding: 1-1                 [300, 20]            [-1, 100, 300]       6,000            6,000
LSTM: 1-2                       --                  [2, 100, 512]        3,768,320        3,760,128
  weight_ih_l0                 [2048, 300]
  weight_hh_l0                 [2048, 512]
  weight_ih_l1                 [2048, 512]
  weight_hh_l1                 [2048, 512]
Linear: 1-3                    [512, 20]            [-1, 100, 20]        10,260           10,240
========================================================================================================
Total params: 3,784,580
Trainable params: 3,784,580
Non-trainable params: 0
--------------------------------------------------------------------------------------------------------
Input size (MB): 0.00
Forward/backward pass size (MB): 1.03
Params size (MB): 14.44
Estimated Total Size (MB): 15.46
--------------------------------------------------------------------------------------------------------

ResNet

import torchvision

model = torchvision.models.resnet50()
summary(model, (3, 224, 224), depth=3)

------------------------------------------------------------------------------------------
Layer (type:depth-idx)                   Output Shape              Param #
==========================================================================================
├─Conv2d: 1-1                            [-1, 64, 112, 112]        9,408
├─BatchNorm2d: 1-2                       [-1, 64, 112, 112]        128
├─ReLU: 1-3                              [-1, 64, 112, 112]        --
├─MaxPool2d: 1-4                         [-1, 64, 56, 56]          --
├─Sequential: 1-5                        [-1, 256, 56, 56]         --
|    └─Bottleneck: 2-1                   [-1, 256, 56, 56]         --
|    |    └─Conv2d: 3-1                  [-1, 64, 56, 56]          4,096
|    |    └─BatchNorm2d: 3-2             [-1, 64, 56, 56]          128
|    |    └─ReLU: 3-3                    [-1, 64, 56, 56]          --
|    |    └─Conv2d: 3-4                  [-1, 64, 56, 56]          36,864
|    |    └─BatchNorm2d: 3-5             [-1, 64, 56, 56]          128
|    |    └─ReLU: 3-6                    [-1, 64, 56, 56]          --
|    |    └─Conv2d: 3-7                  [-1, 256, 56, 56]         16,384
|    |    └─BatchNorm2d: 3-8             [-1, 256, 56, 56]         512
|    |    └─Sequential: 3-9              [-1, 256, 56, 56]         --
|    |    └─ReLU: 3-10                   [-1, 256, 56, 56]         --

  ...
  ...
  ...

├─AdaptiveAvgPool2d: 1-9                 [-1, 2048, 1, 1]          --
├─Linear: 1-10                           [-1, 1000]                2,049,000
==========================================================================================
Total params: 60,192,808
Trainable params: 60,192,808
Non-trainable params: 0
------------------------------------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 344.16
Params size (MB): 229.62
Estimated Total Size (MB): 574.35
------------------------------------------------------------------------------------------

Other Examples

----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1            [-1, 1, 16, 16]              10
              ReLU-2            [-1, 1, 16, 16]               0
            Conv2d-3            [-1, 1, 28, 28]              10
              ReLU-4            [-1, 1, 28, 28]               0
================================================================
Total params: 20
Trainable params: 20
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 0.77
Forward/backward pass size (MB): 0.02
Params size (MB): 0.00
Estimated Total Size (MB): 0.78
----------------------------------------------------------------

References

• Thanks to @sksq96, @nmhkahn, and @sangyx for providing the original code this project was based off of.
• For Model Size Estimation @jacobkimmel (details here)