# deep-learning-teamwork

**Repository Path**: xujijun123/deep-learning-teamwork

## Basic Information

- **Project Name**: deep-learning-teamwork
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: Fcanet
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-05-27
- **Last Updated**: 2024-06-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# README

This is a simple implementation of FCANET, which includes the three types of strategies provided in the essay (https://openaccess.thecvf.com/content/ICCV2021/papers/Qin_FcaNet_Frequency_Channel_Attention_Networks_ICCV_2021_paper.pdf)

##Model Introduction

The models directory has three kinds of models. 

`Resnet.py` is the baseline (the simplified classic Resnet)

`Fcanet.py` is the implementation of FCANET, I write it without other reference but the essay given above.

`source_fcanet.py`is also the implementation of FCANET, I write it but reference this code([cfzd/FcaNet: FcaNet: Frequency Channel Attention Networks (github.com)](https://github.com/cfzd/FcaNet)), I have made a lot of changes and make it more adaptive to my code in Fcanet.py and our tasks.



##Train Options

There are two train codes.`train.py` `source_train.py`



`train.py`is for training the `Resnet.py` and  `Fcanet.py`

You can use `python train.py --help` to see the options

`python train.py --model 1` is for train the  basic Resnet

`python train.py --model 2 --strategy LF` is for train the  FCANET(LF)

The results will be stored in the `results` directory .

You can use `printstructure.py` to print the structure of the network.



source_train.py is for training the `source_fcanet.py`

You can use `python source_train.py --help` to see the options



Thank you for your reading!