# ResNet_modify_35layers

**Repository Path**: hilbert-wang/ResNet_modify_35layers

## Basic Information

- **Project Name**: ResNet_modify_35layers
- **Description**: https://github.com/moddent/ResNet_modify_35layers
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-02-06
- **Last Updated**: 2025-02-06

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# ResNet_35 layers
 Modify ResNet with 35 layers.
 Training on Cifar10.
## Architecture
### Residual block(full pre-activation)
<p align="left">
    <img src="/Residual block.png" width="400" height="680"/>
</p>

### Model Overall
See this [model](https://github.com/moddent/ResNet_customize/blob/main/model.png).


## Training
 Run the following script from the directory:
 
    python cifar10_rec_elu+relu.py
## Result
### Loss
<p align="left">
    <img src="/loss.png" width="640" height="480"/>
</p>

### Accuracy
<p align="left">
    <img src="/acc.png" width="640" height="480"/>
</p>

### Testing

<p align="left">
    <img src="/testing.png" width="600" height="330"/>
</p>