https://github.com/enhuiz/easy-grad-cam
A simple Grad-CAM tool for PyTorch
https://github.com/enhuiz/easy-grad-cam
Last synced: 3 months ago
JSON representation
A simple Grad-CAM tool for PyTorch
- Host: GitHub
- URL: https://github.com/enhuiz/easy-grad-cam
- Owner: enhuiz
- Created: 2020-03-13T07:11:22.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2020-03-13T08:18:46.000Z (about 5 years ago)
- Last Synced: 2025-01-07T18:29:12.422Z (5 months ago)
- Language: Python
- Size: 208 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Easy Grad-CAM: A simple Grad-CAM tool for PyTorch
## Install
```bash
pip install .
```
## Example
```python
import torch
from torchvision import models
import matplotlib.pyplot as plt
from easy_grad_cam import grad_cam, blend
model = models.resnet18(True)
model.eval()
image = plt.imread('cat_dog.png')
image = torch.tensor(image.transpose(2, 0, 1))
# 1. select the conv layer to visualize
conv = list(model.children())[-3]
# 2. create a with statement as this
with grad_cam(conv, size=image.shape[-2:]) as compute:
# 3. forward
logits = model(image[None])
# 4. backward
loss = logits.log_softmax(-1).max(-1)[0]
loss.backward()
# 5. get the cam!
cam = compute()[0]
# 6. blend the cam with the original image
blended = blend(image, cam)
# 7. plot it
plt.imshow(blended)
plt.show()
```
| Original | GradCAM |
| ---------------- | ------------------------ |
|  |  |
## References
Selvaraju, Ramprasaath R., et al. "Grad-cam: Visual explanations from deep networks via gradient-based localization." Proceedings of the IEEE international conference on computer vision. 2017.