Forward pass computation in PyTorch - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of a simple forward pass in PyTorch

What is the output tensor after running this forward pass code?

PyTorch

import torch
import torch.nn as nn

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(2, 1, bias=False)

    def forward(self, x):
        return self.linear(x)

model = SimpleModel()
with torch.no_grad():
    model.linear.weight.copy_(torch.tensor([[2.0, 3.0]]))
    input_tensor = torch.tensor([[1.0, 2.0]])
    output = model(input_tensor)
    print(output)

A tensor([[8.]])

B tensor([[7.]])

C tensor([[5.]])

D tensor([[10.]])

💡 Hint

Remember the forward pass multiplies input by weights and sums them.

The linear layer multiplies input [1, 2] by weights [2, 3] without bias: 1*2 + 2*3 = 8.

❓ Model Choice

intermediate

Choosing the correct model output shape

Given a PyTorch model with nn.Linear(4, 3) and input tensor shape (5, 4), what is the output shape after the forward pass?

A (5, 4)

B (4, 3)

C (3, 5)

D (5, 3)

💡 Hint

The linear layer changes the last dimension from input features to output features.

The input batch size is 5, each with 4 features. The linear layer outputs 3 features per input, so output shape is (5, 3).

❓ Hyperparameter

advanced

Effect of activation function in forward pass

Which activation function will output only values between 0 and 1 during the forward pass?

A ReLU

B Sigmoid

C Tanh

D LeakyReLU

💡 Hint

Think about which function squashes outputs into a probability-like range.

Sigmoid outputs values strictly between 0 and 1, suitable for probabilities.

❓ Metrics

advanced

Calculating accuracy from model predictions

Given model predictions tensor([0, 1, 1, 0]) and true labels tensor([0, 0, 1, 0]), what is the accuracy?

A 0.75

B 0.5

C 0.25

D 1.0

💡 Hint

Accuracy is number of correct predictions divided by total predictions.

Correct predictions are at indices 0, 2, and 3 (3 out of 4), so accuracy is 0.75.

🔧 Debug

expert

Identifying error in forward pass code

What error will this PyTorch forward pass code raise? import torch import torch.nn as nn class Model(nn.Module): def __init__(self): super().__init__() self.linear = nn.Linear(3, 2) def forward(self, x): return self.linear(x) model = Model() input_tensor = torch.randn(4, 4) output = model(input_tensor)

PyTorch

import torch
import torch.nn as nn

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 2)

    def forward(self, x):
        return self.linear(x)

model = Model()
input_tensor = torch.randn(4, 4)
output = model(input_tensor)

A No error, outputs tensor of shape (4, 2)

B TypeError: forward() missing 1 required positional argument

C RuntimeError: size mismatch

D ValueError: input tensor must be 1D

💡 Hint

Check if input tensor shape matches the expected input features of the linear layer.

The linear layer expects input with last dimension 3, but input tensor has last dimension 4, causing size mismatch error.