File size: 1,176 Bytes
817ddf6 82bac57 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 | ---
title: CIFAR10 CustomResnet GradCAM
emoji: ⚡
colorFrom: gray
colorTo: green
sdk: gradio
sdk_version: 3.39.0
app_file: app.py
pinned: false
license: mit
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
# CIFAR10 ResNet18 Gradio Interface
**Training details:**
- This model is trained on CIFAR10 with following classes: ['plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']
**Features**:
- **GradCAM Visualization**: User can adjust the visualization opacity and select model layers to be used.
- **Misclassified Images**: User can view in which images prediction went wrong.
- **Image Upload & Examples**: User can test prediction on new images which can be uploaded.
- **Top Predictions**: User can view model's top class predictions (up to 10).
## **How to Use**:
1. **Choose an Image**: Upload your own or pick from the examples.
2. **Tweak Settings**: Adjust sliders for GradCAM, top predictions, and misclassified images.
3. **Submit**: Click to get results.
**Usecase**: This interface uses a ResNet18 model trained on CIFAR10. GradCAM visualizations highlight the model's focus.
|