PyTorch model input shape

Question

I loaded a custom PyTorch model and I want to find out its input shape. Something like this:

model.input_shape

Is it possible to get this information?

---

Update: print() and summary() don't show this model's input shape, so they are not what I'm looking for.

Norty

hi bro, pytorch model's input shape is flexible. only important thing is its depth, rgb or grayscale. – Kiss

if it's convolutional neural network model.. – Kiss

@yakhyo, so input can be any shape ? – Norty

yes it can be any shape except depth – Kiss

@yakhyo, How it is possible? Could you explain more detailed please – Norty

Could you refer to this – Kiss

Can you provide the model definition? – Block

discuss.pytorch.org/t/cnn-input-image-size-formula/27954/9 – Block

Note: "can be any shape" is technically correct, but of course most networks expect a specific shape (and more important: resolution. That is a dog is expected to be 100px tall rather than 10px). a 10px dog does not produce errors, but won't be detected / classified etc. – Willy

Answer 1

PyTorch flexibility

PyTorch models are very flexible objects, to the point where they do not enforce or generally expect a fixed input shape for data.

If you have certain layers there may be constraints e.g:

• a flatten followed by a fully connected layer of width N would enforce the dimensions of your original input (M1 x M2 x ... Mn) to have a product equal to N
• a 2d convolution of N input channels would enforce the data to be 3 dimensionsal, with the first dimension having size N

But as you can see neither of these enforce the total shape of the data.

We might not realize it right now, but in more complex models, getting the size of the first linear layer right is sometimes a source of frustration. We’ve heard stories of famous practitioners putting in arbitrary numbers and then relying on error messages from PyTorch to backtrack the correct sizes for their linear layers. Lame, eh? Nah, it’s all legit!

• Deep Learning with PyTorch

Investigation

Simple case: First layer is Fully Connected

If your model's first layer is a fully connected one, then the first layer in print(model) will detail the expected dimensionality of a single sample.

Ambiguous case: CNN

If it is a convolutional layer however, since these are dynamic and will stride as long/wide as the input permits, there is no simple way to retrieve this info from the model itself.¹ This flexibility means that for many architectures multiple compatible input sizes² will all be acceptable by the network.

This is a feature of PyTorch's Dynamic computational graph.

Manual inspection

What you will need to do is investigate the network architecture, and once you've found an interpretable layer (if one is present e.g. fully connected) "work backwards" with its dimensions, determining how the previous layers (e.g. poolings and convolutions) have compressed/modified it.

Example

e.g. in the following model from Deep Learning with PyTorch (8.5.1):

class NetWidth(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 32, kernel_size=3, padding=1)
        self.conv2 = nn.Conv2d(32, 16, kernel_size=3, padding=1)
        self.fc1 = nn.Linear(16 * 8 * 8, 32)
        self.fc2 = nn.Linear(32, 2)
    
    def forward(self, x):
        out = F.max_pool2d(torch.tanh(self.conv1(x)), 2)
        out = F.max_pool2d(torch.tanh(self.conv2(out)), 2)
        out = out.view(-1, 16 * 8 * 8)
        out = torch.tanh(self.fc1(out))
        out = self.fc2(out)
        return out

We see the model takes an input 2.d. image with 3 channels and:

• Conv2d -> sends it to an image of the same size with 32 channels
• max_pool2d(,2) -> halves the size of the image in each dimension
• Conv2d -> sends it to an image of the same size with 16 channels
• max_pool2d(,2) -> halves the size of the image in each dimension
• view -> reshapes the image
• Linear -> takes a tensor of size 16 * 8 * 8 and sends to size 32
• ...

So working backwards, we have:

• a tensor of shape 16 * 8 * 8
• un-reshaped into shape (channels x height x width)
• un-max_pooled in 2d with factor 2, so height and width un-halved
• un-convolved from 16 channels to 32
  Hypothesis: It is likely 16 in the product thus refers to the number of channels, and that the image seen by view was of shape (channels, 8,8), and currently is (channels, 16,16)²
• un-max_pooled in 2d with factor 2, so height and width un-halved again (channels, 32,32)
• un-convolved from 32 channels to 3

So assuming the kernel_size and padding are sufficient that the convolutions themselves maintain image dimensions, it is likely that the input image is of shape (3,32,32) i.e. RGB 32x32 pixel square images.

---

Notes:

1. Even the external package pytorch-summary requires you provide the input shape in order to display the shape of the output of each layer.

2. It could however be any 2 numbers whose produce equals 8*8 e.g. (64,1), (32,2), (16,4) etc however since the code is written as 8*8 it is likely the authors used the actual dimensions.

Block

Answer 2

print(model)

Will give you a summary of the model, where you can see the shape of each layer.

You can also use the pytorch-summary package.

If your network has a FC as a first layer, you can easily figure its input shape. You mention that you have a Convolutional layer at the front. With Fully Connected layers present too, the network will produce output for only one specific input size. I'm proposing to figure this out by using various shapes, i.e. feeding a toy batch with some shape, and then checking the output of the Conv layer just before the FC layer.

As this depends on the architecture of the net before the first FC layer (num of conv layers, kernels, etc), I can't give you an exact formula for the correct input. As mentioned, you have to figure this out by experimenting with various input shapes, and the resulting net's output before the first FC. There's (almost) always a way to solve something with code, but I can't think of something else right now.

Yeseniayeshiva

but it's not about input_shape – Kiss

You can make out input shape from the first layer. If the first layer is a convnet, you can evaluate if an input will 'crash' the code during inference at the first FC layer. – Yeseniayeshiva

according to pytorch's doc, there are only in_channels and out_channels that you need to specify. – Kiss

Yes, that's it for defining a conv net. So? – Yeseniayeshiva

@AlexMetsai, yes my model is CNN. So, how should I find out the input shape? What do you mean by "evaluate if an input will 'crash' the code"? I should keep guessing ??? – Norty

This answer is off topic, print() and pytorch-summery don't show input shape. They show every layer's output shape. – Norty

If your network has a FC as a first layer, you can easily figure the its shape. You mention that you have a Convolutional layer at the front. If you are using Fully Connected layers, which I'm guessing that you do, the network will produce output for only one specific input size. I'm proposing to figure this out with various shapes, e.g. feeding a toy batch with some shape, and check the output of the Conv layer just before the FC layer. As this depends of the architecture of the net before the first FC layer (num of conv layers, kernels, etc), I can't give you an exact formula. – Yeseniayeshiva

As mentioned in the other answer: "We’ve heard stories of famous practitioners putting in arbitrary numbers and then relying on error messages from PyTorch to backtrack the correct sizes for their linear layers. Lame, eh? Nah, it’s all legit!" – Yeseniayeshiva

Answer 3

You can get input shape from first tensor in model parameters.

For example create some model:

class CustomNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc1 = nn.Linear(1568, 256)
        self.fc2 = nn.Linear(256, 256)
        self.fc3 = nn.Linear(256, 20)

    def forward(self, x):
        out = self.fc1(x)
        out = F.relu(out)
        out = self.fc2(out)
        out = F.relu(out)
        out = self.fc3(out)
        return out

model = CustomNet()

So model.parameters() method returns an iterator over module parameters of torch.Tensor class. Look at the docs https://pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.parameters

And first parameter is the input tensor.

first_parameter = next(model.parameters())
input_shape = first_parameter.size()

Sumo

Hi Alexander! 1. Your code has a typo in first_parameter = next(module.parameters()). It should be --> model.parameters(). 2. Seems like it works if network starts with FC layer. – Hosfmann