How to beginner · 3 min read

How to use MSELoss in PyTorch

Q: How to use MSELoss in PyTorch

Use torch.nn.MSELoss in PyTorch to compute the mean squared error between predicted and target tensors. Instantiate it as a loss function, then call it with model outputs and ground truth labels to get the loss value.

Quick answer

Use torch.nn.MSELoss in PyTorch to compute the mean squared error between predicted and target tensors. Instantiate it as a loss function, then call it with model outputs and ground truth labels to get the loss value.

PREREQUISITES

Python 3.8+
pip install torch>=2.0

Setup

Install PyTorch if you haven't already. Use the official command from PyTorch installation guide. For CPU-only:

pip install torch torchvision torchaudio

bash

pip install torch torchvision torchaudio

Step by step

This example shows how to use torch.nn.MSELoss to compute the mean squared error loss between predictions and targets in a regression setting.

python

import torch
import torch.nn as nn

# Create dummy prediction and target tensors
predictions = torch.tensor([[0.5], [0.8], [0.3]], dtype=torch.float32)
targets = torch.tensor([[0.0], [1.0], [0.0]], dtype=torch.float32)

# Instantiate MSELoss
criterion = nn.MSELoss()

# Compute loss
loss = criterion(predictions, targets)
print(f'MSE Loss: {loss.item():.4f}')

output

MSE Loss: 0.0867

Common variations

You can use reduction parameter in MSELoss to control how the loss is aggregated: 'mean' (default), 'sum', or 'none' to get element-wise loss. Also, use it in training loops with optimizers.

python

import torch
import torch.nn as nn

# Example with reduction='sum'
criterion_sum = nn.MSELoss(reduction='sum')
loss_sum = criterion_sum(predictions, targets)
print(f'Sum reduction MSE Loss: {loss_sum.item():.4f}')

# Example with reduction='none'
criterion_none = nn.MSELoss(reduction='none')
loss_none = criterion_none(predictions, targets)
print(f'Element-wise MSE Loss:\n{loss_none}')

output

Sum reduction MSE Loss: 0.2600
Element-wise MSE Loss:
tensor([[0.2500],
        [0.0400],
        [0.0900]])

Troubleshooting

If you get a shape mismatch error, ensure predictions and targets have the same shape and dtype.
Use predictions.float() and targets.float() if needed.
For GPU usage, move tensors and loss function to the same device.

✅

Key Takeaways

Use torch.nn.MSELoss() to compute mean squared error between predictions and targets.
The reduction parameter controls how losses are aggregated: mean, sum, or none.
Ensure predictions and targets have matching shapes and data types to avoid errors.
MSELoss is commonly used for regression problems in PyTorch models.

Verified 2026-04

Verify ↗