KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #claude-code #codex #dsh #dsh-plugin #ai-agent #ai-agents #deepseek-harness #agent-skills #macos #cli

GAN-Metrics

★ 60 Open GitHub ↗

A collection of metrics for evaluating GAN models.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Metrics of GANs
 
[![Maintenance](https://img.shields.io/badge/Maintaining%3F-yes-green.svg)](https://raw.githubusercontent.com/yhlleo/GAN-Metrics/master/(https:/github.com/yhlleo/frechet-bert-distance/graphs/commit-activity))
![Contributing](https://img.shields.io/badge/contributions-welcome-brightgreen.svg?style=flat)

- [x] Fréchet Inception Distance (**FID**)
  - [x] Inception Score (**IS**)
  - [x] Number of statistically-Different Bins (**NDB**) 
  - [x] Jensen-Shannon Divergence (**JSD**)
  - [x] Learned Perceptual Image Patch Similarity (**LPIPS**)

## Configuration

See the [`environment.yaml`](./environment.yaml). We provide an user-friendly configuring method via Conda system, and you can create a new Conda environment using the command:

```
conda env create -f environment.yaml
```

## Usage

- IS:
```
python eval.py --metric is --pred_list <path/to/pred_list> --gpu_id 0 --resize 299
```

- FID:

```
python eval.py --metric fid --pred_list <path/to/pred_list> --gt_list <path/to/gt_list> --gpu_id 0 --resize 299
```

- NBD & JSD:
```
python eval.py --metric ndb --pred_list <path/to/pred_list> --gt_list <path/to/gt_list> --gpu_id 0 --resize 128
```

- LPIPS:

```
python lpips.py --path <path/to/image_folder> --test_list <path/to/test_list>
```

Example of test list in LPIPS:

```
a1.png	a2.png
a1.png	a3.png
...
```

GAN-Metrics

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README