fix bus and improve README

.gitignore

@@ -9,4 +9,5 @@ __pycache__
 .python-version
 
 build
-dist
+dist
+pytorch_gan_metrics.egg-info

README.md

@@ -5,38 +5,44 @@
 ## Notes
 The FID implementation is inspired from [pytorch-fid](https://github.com/mseitzer/pytorch-fid).
 
+This repository is developed for personal research. If you think this package can also benefit your life, please feel free to open issues.
+
+## Install
+```
+pip install pytorch-gan-metrics
+```
+
 ## Feature
 - Currently, this package supports following metrics:
   - [Inception Score](https://github.com/openai/improved-gan) (IS)
   - [Fréchet Inception Distance](https://github.com/bioinf-jku/TTUR) (FID)
 - The computation processes of IS and FID are integrated to avoid multiple forward propagations.
-- Read image on the fly for both metrics.
+- Support reading image on the fly to avoid out of memory especially for large scale images.
+- Support computation on GPU to speed up some cpu operations such as `np.cov` and `scipy.linalg.sqrtm`.
 
-## Reproducing Results of Official Implementations
+## Reproducing Results of Official Implementations on CIFAR-10
 
-- CIFAR-10
-
-    |                   |Train IS  |Test IS   |Train(50k) vs Test(10k)<br>FID|
-    |-------------------|:--------:|:--------:|:----------------------------:|
-    |Official           |11.24±0.20|10.98±0.22|3.1508                        |
-    |pytorch-gan-metrics|11.26±0.27|10.97±0.33|3.1517                        |
+|                   |Train IS  |Test IS   |Train(50k) vs Test(10k)<br>FID|
+|-------------------|:--------:|:--------:|:----------------------------:|
+|Official           |11.24±0.20|10.98±0.22|3.1508                        |
+|pytorch-gan-metrics|11.26±0.27|10.97±0.33|3.1517                        |
+|pytorch-gan-metrics<br>`use_torch=True`|11.26±0.21|10.97±0.34|3.1377                        |
 
-    Due to the framework difference between PyTorch and TensorFlow, the results are slightly different from official implementations.
-
-## Install
-```
-pip install pytorch-gan-metrics
-```
+The results are slightly different from official implementations due to the framework difference between PyTorch and TensorFlow.
 
 ## Prepare Statistics for FID
-- [Download](https://drive.google.com/drive/folders/1UBdzl6GtNMwNQ5U-4ESlIer43tNjiGJC?usp=sharing) precalculated statistics for dataset or
+- [Download](https://drive.google.com/drive/folders/1UBdzl6GtNMwNQ5U-4ESlIer43tNjiGJC?usp=sharing) precalculated statistics or
 - Calculate statistics for your custom dataset using command line tool
     ```bash
     python -m pytorch_gan_metrics.calc_fid_stats --path path/to/images --output name.npz
     ```
     See [calc_fid_stats.py](./pytorch_gan_metrics/calc_fid_stats.py) for implementation details.
 
 ### Documentation
+
+#### How to use GPU?
+`pytorch_gan_metrics` default uses `torch.device('cuda:0')` if GPU is available; Otherwise, it uses `cpu` to calculate inception feature.
+
 #### Using `torch.Tensor` as images
 - Prepare images in type `torch.float32` with shape `[N, 3, H, W]` and normalized to `[0,1]`.
     ```python
@@ -54,8 +60,9 @@ pip install pytorch-gan-metrics
         images, 'path/to/statistics.npz')
 
     ```
-#### Using PyTorch DataLoader
-- Use `pytorch_gan_metrics.ImageDataset` to collect images on disk or use custom dataset which should only return an image in `__getitem__`.
+
+#### Using PyTorch DataLoader to Provide Images
+- Use `pytorch_gan_metrics.ImageDataset` to collect images on disk or use custom `torch.utils.data.Dataset` which should only return an image in the end of `__getitem__`.
     ```python
     from pytorch_gan_metrics import ImageDataset
 
@@ -92,7 +99,7 @@ pip install pytorch-gan-metrics
         loader, 'path/to/statistics.npz')
     ```
 
-#### From directory
+#### Specify Images by a Directory Path
 - Calculate metrics for images in the directory.
     ```python
     from pytorch_gan_metrics import (
@@ -106,6 +113,11 @@ pip install pytorch-gan-metrics
         'path/to/images', fid_stats_path)
     ```
 
+#### Set PyTorch as backend
+- Set `use_torch=True` when calling functions `get_*` such as `get_inception_score`, `get_fid`, etc.
+- **WARNING** when set `use_torch=True`, the FID might be `nan` due to the unstable implementation of matrix sqrt.
+- This option is recommended to be used when evaluate generative models on a server machine which is equipped with high efficiency GPUs while the cpu frequency is low.
+
 ## License
 
 This implementation is licensed under the Apache License 2.0.

pytorch_gan_metrics/__init__.py

@@ -17,4 +17,4 @@
     get_inception_score_and_fid_from_directory
 ]
 
-__version__ = '0.1.0'
+__version__ = '0.2.0'

pytorch_gan_metrics/calc_fid_stats.py

@@ -16,8 +16,6 @@
                         help="output path")
     parser.add_argument("--batch_size", type=int, default=50,
                         help="batch size (default=50)")
-    parser.add_argument("--inception_dir", type=str, default='/tmp',
-                        help='path to inception model dir')
     args = parser.parse_args()
 
     dataset = ImageDataset(args.path, exts=['png', 'jpg'])

pytorch_gan_metrics/calc_metrics.py

@@ -16,5 +16,5 @@
     dataset = ImageDataset(args.path, exts=['png', 'jpg'])
     loader = DataLoader(dataset, batch_size=50, num_workers=4)
     (IS, IS_std), FID = get_inception_score_and_fid(
-        loader, args.stats, use_torch=True, verbose=True)
+        loader, args.stats, verbose=True)
     print(IS, IS_std, FID)

pytorch_gan_metrics/inception.py

@@ -6,8 +6,8 @@
 
 # Inception weights ported to Pytorch from
 # http://download.tensorflow.org/models/image/imagenet/inception-2015-12-05.tgz
-FID_WEIGHTS_URL = ('https://github.com/mseitzer/pytorch-fid/releases/download/'
-                   'fid_weights/pt_inception-2015-12-05-6726825d.pth')
+FID_WEIGHTS_URL = ('https://github.com/w86763777/pytorch-gan-metrics/releases/'
+                   'download/v0.1.0/pt_inception-2015-12-05-6726825d.pth')
 
 
 class InceptionV3(nn.Module):

setup.py

@@ -1,8 +1,10 @@
 import os
-
 import setuptools
 
 
+import pytorch_gan_metrics
+
+
 def read(rel_path):
     base_path = os.path.abspath(os.path.dirname(__file__))
     with open(os.path.join(base_path, rel_path), 'r') as f:
@@ -12,20 +14,21 @@ def read(rel_path):
 if __name__ == '__main__':
     setuptools.setup(
         name='pytorch_gan_metrics',
-        version='0.1.0',
+        version=pytorch_gan_metrics.__version__,
         author='Yi-Lun Wu',
         author_email='[email protected]',
-        description=(
-            'Package for calculating GAN metrics using Pytorch'),
+        description=('Package for calculating GAN metrics using Pytorch'),
         long_description=read('README.md'),
         long_description_content_type='text/markdown',
         url='https://github.com/w86763777/pytorch-gan-metrics',
         packages=setuptools.find_packages(include=['pytorch_gan_metrics']),
         keywords=[
             'PyTorch',
             'GAN',
-            'Inception Score', 'IS',
-            'Frechet Inception Distance', 'FID'],
+            'Inception Score',
+            'IS',
+            'Frechet Inception Distance',
+            'FID'],
         classifiers=[
             'Programming Language :: Python :: 3',
             'License :: OSI Approved :: Apache Software License',