# Paddle-Image-Models

**Repository Path**: AgentMaker/Paddle-Image-Models

## Basic Information

- **Project Name**: Paddle-Image-Models
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: dev
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-05-11
- **Last Updated**: 2021-05-11

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Paddle-Image-Models
![GitHub forks](https://img.shields.io/github/forks/AgentMaker/Paddle-Image-Models)
![GitHub Repo stars](https://img.shields.io/github/stars/AgentMaker/Paddle-Image-Models)
![GitHub release (latest by date including pre-releases)](https://img.shields.io/github/v/release/AgentMaker/Paddle-Image-Models?include_prereleases)
![GitHub](https://img.shields.io/github/license/AgentMaker/Paddle-Image-Models)  

[English](README.md) | 简体中文

一个基于飞桨框架实现的图像预训练模型库。

![](https://ai-studio-static-online.cdn.bcebos.com/34e7bbbc80d24412b3c21efb56778ad43b53f9b1be104e499e0ff8b663a64a53)

## 安装
* 通过 pip 进行安装：

    ```shell
    $ pip install ppim
    ```

* 通过 whl 包进行安装：[【Releases Packages】](https://github.com/AgentMaker/Paddle-Image-Models/releases)

## 使用方法
* 快速使用

    ```python
    import paddle
    from ppim import rednet_26

    # 加载模型
    model, val_transforms = rednet_26(pretrained=True)

    # 模型结构总览 
    paddle.summary(model, input_size=(1, 3, 224, 224))

    # 准备一个随机的输入
    x = paddle.randn(shape=(1, 3, 224, 224))

    # 模型前向计算
    out = model(x)
    ```

* 模型微调
    
    ```python
    import paddle
    import paddle.nn as nn
    import paddle.vision.transforms as T
    from paddle.vision import Cifar100

    from ppim import rexnet_1_0

    # 加载模型
    model, val_transforms = rexnet_1_0(pretrained=True)

    # 使用飞桨高层 API Model
    model = paddle.Model(model)

    # 配置优化器
    opt = paddle.optimizer.Adam(learning_rate=0.001, parameters=model.parameters())

    # 配置损失函数
    loss = nn.CrossEntropyLoss()

    # 配置评估指标
    metric = paddle.metric.Accuracy(topk=(1, 5))

    # 模型准备
    model.prepare(optimizer=opt, loss=loss, metrics=metric)

    # 配置训练集数据处理
    train_transforms = T.Compose([
        T.Resize(256, interpolation='bicubic'),
        T.RandomCrop(224),
        T.ToTensor(),
        T.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
    ])

    # 加载 Cifar100 数据集
    train_dataset = Cifar100(mode='train', transform=train_transforms, backend='pil')
    val_dataset = Cifar100(mode='test',  transform=val_transforms, backend='pil')

    # 模型微调
    model.fit(
        train_data=train_dataset, 
        eval_data=val_dataset, 
        batch_size=256, 
        epochs=2, 
        eval_freq=1, 
        log_freq=1, 
        save_dir='save_models', 
        save_freq=1, 
        verbose=1, 
        drop_last=False, 
        shuffle=True,
        num_workers=0
    )
    ```

## 模型库

* [DLA](./docs/cn/model_zoo/dla.md)

* [ReXNet](./docs/cn/model_zoo/rexnet.md)

* [RedNet](./docs/cn/model_zoo/rednet.md)

* [RepVGG](./docs/cn/model_zoo/repvgg.md)

* [HarDNet](./docs/cn/model_zoo/hardnet.md)

* [PiT](./docs/cn/model_zoo/pit.md)

* [PVT](./docs/cn/model_zoo/pvt.md)

* [TNT](./docs/cn/model_zoo/tnt.md)

* [DeiT](./docs/cn/model_zoo/deit.md)