https://github.com/enochkan/torch-metrics

Metrics for model evaluation in pytorch
https://github.com/enochkan/torch-metrics

computer-vision deep-learning machine-learning metrics pytorch pytorch-cnn pytorch-implementation torch-metrics

Last synced: about 1 month ago
JSON representation

Metrics for model evaluation in pytorch

• Host: GitHub
• URL: https://github.com/enochkan/torch-metrics
• Owner: enochkan
• License: mit
• Created: 2020-11-03T19:08:46.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2021-03-22T18:02:50.000Z (about 3 years ago)
• Last Synced: 2024-04-16T07:48:57.237Z (about 1 month ago)
• Topics: computer-vision, deep-learning, machine-learning, metrics, pytorch, pytorch-cnn, pytorch-implementation, torch-metrics
• Language: Python
• Homepage:
• Size: 42 KB
• Stars: 110
• Watchers: 3
• Forks: 14
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• Awesome-pytorch-list - torch-metrics
• Awesome-pytorch-list-CNVersion - torch-metrics

README

        
# Torch-metrics

[![PyPI version](https://badge.fury.io/py/torch-metrics.svg)](https://badge.fury.io/py/torch-metrics)

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

Model evaluation metrics for PyTorch

Torch-metrics serves as a custom library to provide common ML evaluation metrics in Pytorch, similar to `tf.keras.metrics`.

As summarized in this [issue](https://github.com/pytorch/pytorch/issues/22439), Pytorch does not have a built-in libary `torch.metrics` for model evaluation metrics. This is similar to the metrics library in [PyTorch Lightning](https://pytorch-lightning.readthedocs.io/en/latest/metrics.html#class-metrics).

## Usage

- `pip install --upgrade torch-metrics`

```python

from torch_metrics import Accuracy

## define metric ##

metric = Accuracy(from_logits=False)

y_pred = torch.tensor([1, 2, 3, 4])

y_true = torch.tensor([0, 2, 3, 4])

print(metric(y_pred, y_true))

```

```python

## define metric ##

metric = Accuracy()

y_pred = torch.tensor([[0.2, 0.6, 0.1, 0.05, 0.05],

                       [0.2, 0.1, 0.6, 0.05, 0.05],

                       [0.2, 0.05, 0.1, 0.6, 0.05],

                       [0.2, 0.05, 0.05, 0.05, 0.65]])

y_true = torch.tensor([0, 2, 3, 4])

print(metric(y_pred, y_true))

```

## Implementation

Metrics from `tf.keras.metrics` and other metrics that are already implemented vs to-do

- [x] MeanSquaredError class

- [x] RootMeanSquaredError class

- [x] MeanAbsoluteError class

- [x] Precision class

- [x] Recall class

- [x] MeanIoU class

- [x] DSC class (Dice Similarity Coefficient)

- [x] F1Score class

- [x] RSquared class

- [x] Hinge class

- [x] SquaredHinge class

- [x] LogCoshError class

- [x] Accuracy class

- [x] KLDivergence class

- [ ] CosineSimilarity class

- [ ] AUC class

- [ ] BinaryCrossEntropy class

- [ ] CategoricalCrossEntropy class

- [ ] SparseCategoricalCrossentropy class

## Local Development and Testing

To quickly get started with local development, run:

```python

make develop

```

To test, run:

```python

python3 -m pytest

```

### Pre-commit hooks

To run pre-commit against all files:

```python

pre-commit run --all-files

```

## Contributing

Please raise issues or feature requests [here](https://github.com/enochkan/torch-metrics/issues). It will be extremely helpful if you comment on a specific issue before working on it. This provides visibility for others who also intend to work on the same issue. Reference any pull requests to their original issues.