cuicheng01
GitHub
parent
10cf2775e4
commit
43e6382aa3
|
|
@ -0,0 +1,315 @@
|
|||
# PP-HGNeV2 系列
|
||||
---
|
||||
- [1. 模型介绍](#1)
|
||||
- [1.1 模型简介](#1.1)
|
||||
- [1.2 模型细节](#1.2)
|
||||
- [1.3 实验精度](#1.3)
|
||||
- [2. 模型训练、评估和预测](#2)
|
||||
- [2.1 环境配置](#2.1)
|
||||
- [2.2 数据准备](#2.2)
|
||||
- [2.3 模型训练](#2.3)
|
||||
- [2.3.1 从头训练 ImageNet](#2.3.1)
|
||||
- [2.3.2 基于 ImageNet 权重微调其他分类任务](#2.3.2)
|
||||
- [2.4 模型评估](#2.4)
|
||||
- [2.5 模型预测](#2.5)
|
||||
- [3. 模型推理部署](#3)
|
||||
- [3.1 推理模型准备](#3.1)
|
||||
- [3.2 基于 Python 预测引擎推理](#3.2)
|
||||
- [3.2.1 预测单张图像](#3.2.1)
|
||||
- [3.2.2 基于文件夹的批量预测](#3.2.2)
|
||||
- [3.3 基于 C++ 预测引擎推理](#3.3)
|
||||
- [3.4 服务化部署](#3.4)
|
||||
- [3.5 端侧部署](#3.5)
|
||||
- [3.6 Paddle2ONNX 模型转换与预测](#3.6)
|
||||
|
||||
<a name='1'></a>
|
||||
|
||||
## 1. 模型介绍
|
||||
|
||||
<a name='1.1'></a>
|
||||
|
||||
### 1.1 模型简介
|
||||
|
||||
PP-HGNetV2(High Performance GPU Network V2) 是百度飞桨视觉团队自研的 PP-HGNet 的下一代版本，其在 PP-HGNet 的基础上，做了进一步优化和改进，最终在 NVIDIA GPU 设备上，将 "Accuracy-Latency Balance" 做到了极致，精度大幅超过了其他同样推理速度的模型。其在单标签分类、多标签分类、目标检测、语义分割等任务中，均有较强的表现。
|
||||
|
||||
<a name='1.2'></a>
|
||||
|
||||
### 1.2 模型细节
|
||||
|
||||
PP-HGNetV2 在 PP-HGNet 上的具体改进点如下：
|
||||
|
||||
- 改进了 PPHGNet 网络 stem 部分，堆叠更多的 2x2 卷积核以学习更丰富的局部特征，使用更小的通道数以提升大分辨率任务如目标检测、语义分割等的推理速度；
|
||||
- 替换了 PP-HGNet 中靠后 stage 的较冗余的标准卷积层为 PW + DW5x5 组合，在获得更大感受野的同时网络的参数量更少，且精度可以进一步提升；
|
||||
- 增加了 LearnableAffineBlock 模块，其可以在增加极少参数量的同时大幅提升较小模型的精度，且对推理时间无损；
|
||||
- 重构了 PP-HGNet 网络的 stage 分布，使其涵盖了从 B0-B7 不同量级的模型，从而满足不同任务的需求。
|
||||
|
||||
除以上改进点之外，相比 PaddleClas 提供的其他模型，PP-HGNetV2 默认提供了精度更高、泛化能力更强的 [SSLD](https://arxiv.org/abs/2103.05959) 预训练权重，其在下游任务中表现更佳。
|
||||
|
||||
<a name='1.3'></a>
|
||||
|
||||
### 1.3 模型精度
|
||||
|
||||
PP-HGNetV2 的精度、速度指标、预训练权重、推理模型权重链接如下：
|
||||
|
||||
| Model | Top-1 Acc(\%)(stage-2) | Top-5 Acc(\%)(stage-2) | Latency(ms) | stage-1预训练模型下载地址 | stage-2预训练模型下载地址 |inference模型下载地址(stage-2) |
|
||||
|:--: |:--: |:--: |:--: | :--: |:--: |:--: |
|
||||
| PPHGNetV2_B0 | 77.77 | 93.91 | 0.52 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B0_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B0_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B0_ssld_infer.tar) |
|
||||
| PPHGNetV2_B1 | 79.18 | 94.57 | 0.58 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B1_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B1_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B1_ssld_infer.tar) |
|
||||
| PPHGNetV2_B2 | 81.74 | 95.88 | 0.95 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B2_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B2_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B2_ssld_infer.tar) |
|
||||
| PPHGNetV2_B3 | 82.98 | 96.43 | 1.18 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B3_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B3_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B3_ssld_infer.tar) |
|
||||
| PPHGNetV2_B4 | 83.57 | 96.72 | 1.46 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B4_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B4_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B4_ssld_infer.tar) |
|
||||
| PPHGNetV2_B5 | 84.75 | 97.32 | 2.84 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B5_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B5_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B5_ssld_infer.tar) |
|
||||
| PPHGNetV2_B6 | 86.30 | 97.84 | 5.29 |[下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B6_ssld_stage1_pretrained.pdparams)| [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B6_ssld_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/PPHGNetV2_B6_ssld_infer.tar) |
|
||||
| PPHGNetV2_B7 | comming soon | comming soon | 11.06 |comming soon| comming soon | comming soon |
|
||||
|
||||
|
||||
**备注：**
|
||||
|
||||
* 测试环境：V100，FP32+TensorRT8.5，BS=1；
|
||||
* 为了让下游任务有更高的精度，PP-HGNetV2 全系列提供了 `SSLD` 预训练权重。关于 `SSLD` 相关的内容介绍和训练方法，可以查看[SSLD paper](https://arxiv.org/abs/2103.05959)、[SSLD 训练](../../training/advanced/knowledge_distillation.md)，此处提供的 stage-1 的权重为 `SSLD` 的 stage-1 阶段使用 ImageNet1k+ImageNet22k 挖掘数据蒸馏训练得到的权重，stage-2 权重为 `SSLD` 的 stage-2 阶段使用 ImageNet1k 蒸馏微调得到的权重。在实际场景中，stage-1 的权重有更好的泛化性，建议直接使用 stage-1 的权重来做下游任务训练。
|
||||
|
||||
|
||||
<a name="2"></a>
|
||||
|
||||
## 2. 模型训练、评估和预测
|
||||
|
||||
<a name="2.1"></a>
|
||||
|
||||
### 2.1 环境配置
|
||||
|
||||
* 安装：请先参考文档[环境准备](../../installation.md) 配置 PaddleClas 运行环境。
|
||||
|
||||
<a name="2.2"></a>
|
||||
|
||||
### 2.2 数据准备
|
||||
|
||||
请在[ImageNet 官网](https://www.image-net.org/)准备 ImageNet-1k 相关的数据。
|
||||
|
||||
|
||||
进入 PaddleClas 目录。
|
||||
|
||||
```
|
||||
cd path_to_PaddleClas
|
||||
```
|
||||
|
||||
进入 `dataset/` 目录，将下载好的数据命名为 `ILSVRC2012` ，存放于此。 `ILSVRC2012` 目录中具有以下数据：
|
||||
|
||||
```
|
||||
├── train
|
||||
│ ├── n01440764
|
||||
│ │ ├── n01440764_10026.JPEG
|
||||
│ │ ├── n01440764_10027.JPEG
|
||||
├── train_list.txt
|
||||
...
|
||||
├── val
|
||||
│ ├── ILSVRC2012_val_00000001.JPEG
|
||||
│ ├── ILSVRC2012_val_00000002.JPEG
|
||||
├── val_list.txt
|
||||
```
|
||||
|
||||
其中 `train/` 和 `val/` 分别为训练集和验证集。`train_list.txt` 和 `val_list.txt` 分别为训练集和验证集的标签文件。
|
||||
|
||||
**备注：**
|
||||
|
||||
* 关于 `train_list.txt`、`val_list.txt`的格式说明，可以参考[PaddleClas分类数据集格式说明](../../training/single_label_classification/dataset.md#1-数据集格式说明) 。
|
||||
|
||||
|
||||
<a name="2.3"></a>
|
||||
|
||||
### 2.3 模型训练
|
||||
|
||||
<a name="2.3.1"></a>
|
||||
|
||||
#### 2.3.1 从头训练 ImageNet
|
||||
|
||||
在 `ppcls/configs/ImageNet/PPHGNetV2/` 中提供了 PPHGNetV2 不同大小模型的训练配置，可以加载对应模型的配置训练。如训练 `PPHGNetV2_B4`，则可以通过如下脚本启动训练：
|
||||
|
||||
```shell
|
||||
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
|
||||
python -m paddle.distributed.launch \
|
||||
--gpus="0,1,2,3,4,5,6,7" \
|
||||
tools/train.py \
|
||||
-c ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4.yaml \
|
||||
-o Global.output_dir=./output/PPHGNetV2_B4 \
|
||||
-o Arch.pretrained=False
|
||||
```
|
||||
|
||||
|
||||
**备注：**
|
||||
|
||||
* 当前精度最佳的模型会保存在 `output/PPHGNetV2_B4/best_model.pdparams`;
|
||||
* 此处只是展示了如何从头训练 ImageNet数据，该配置并未使用激进的训练策略或者蒸馏训练策略，所以训练得到的精度较 [1.3](#1.3) 小节要低。如果希望得到 [1.3](#1.3) 小节中的精度，可以查看[SSLD 训练](../../training/advanced/knowledge_distillation.md)，配置好相关的数据，加载 [stage-1 配置](../../../../ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4_ssld_stage1.yaml)、[stage-2 配置](../../../../PPHGNetV2_B4_ssld_stage2.yaml)训练即可。
|
||||
|
||||
<a name="2.3.2"></a>
|
||||
|
||||
#### 2.3.2 基于 ImageNet 权重微调其他分类任务
|
||||
|
||||
模型微调时，需要加载预训练权重，同时需要缩小学习率，以免破坏原有权重。如微调训练 `PPHGNetV2_B4`，则可以通过如下脚本启动训练：
|
||||
|
||||
```shell
|
||||
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
|
||||
python -m paddle.distributed.launch \
|
||||
--gpus="0,1,2,3,4,5,6,7" \
|
||||
tools/train.py \
|
||||
-c ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4.yaml \
|
||||
-o Global.epochs=30 \
|
||||
-o Global.output_dir=./output/PPHGNetV2_B4 \
|
||||
-o Optimizer.lr.learning_rate=0.05
|
||||
|
||||
```
|
||||
**备注：**
|
||||
|
||||
* `epochs` 和 `learning_rate` 可以根据实际情况调整；
|
||||
* 为了更好的泛化性，此处默认加载的权重为 `SSLD` stage-1 训练得到的权重。
|
||||
|
||||
<a name="2.4"></a>
|
||||
|
||||
### 2.4 模型评估
|
||||
|
||||
训练好模型之后，可以通过以下命令实现对模型指标的评估。
|
||||
|
||||
```shell
|
||||
python tools/eval.py \
|
||||
-c ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4.yaml \
|
||||
-o Global.pretrained_model=output/PPHGNetV2_B4/best_model
|
||||
```
|
||||
|
||||
其中 `-o Global.pretrained_model="output/PPHGNetV2_B4/best_model"` 指定了当前最佳权重所在的路径，如果指定其他权重，只需替换对应的路径即可。
|
||||
|
||||
<a name="2.5"></a>
|
||||
|
||||
### 2.5 模型预测
|
||||
|
||||
模型训练完成之后，可以加载训练得到的预训练模型，进行模型预测。在模型库的 `tools/infer.py` 中提供了完整的示例，只需执行下述命令即可完成模型预测：
|
||||
|
||||
```shell
|
||||
python tools/infer.py \
|
||||
-c ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4.yaml \
|
||||
-o Global.pretrained_model=output/PPHGNetV2_B4/best_model
|
||||
```
|
||||
|
||||
输出结果如下：
|
||||
|
||||
```
|
||||
[{'class_ids': [8, 7, 86, 82, 83], 'scores': [0.92473, 0.07478, 0.00025, 7e-05, 6e-05], 'file_name': 'docs/images/inference_deployment/whl_demo.jpg', 'label_names': ['hen', 'cock', 'partridge', 'ruffed grouse, partridge, Bonasa umbellus', 'prairie chicken, prairie grouse, prairie fowl']}]
|
||||
```
|
||||
|
||||
**备注：**
|
||||
|
||||
* 这里`-o Global.pretrained_model="output/PPHGNetV2_B4/best_model"` 指定了当前最佳权重所在的路径，如果指定其他权重，只需替换对应的路径即可。
|
||||
|
||||
* 默认是对 `docs/images/inference_deployment/whl_demo.jpg` 进行预测，此处也可以通过增加字段 `-o Infer.infer_imgs=xxx` 对其他图片预测。
|
||||
|
||||
* 默认输出的是 Top-5 的值，如果希望输出 Top-k 的值，可以指定`-o Infer.PostProcess.topk=k`，其中，`k` 为您指定的值。
|
||||
|
||||
* 默认的标签映射基于 ImageNet 数据集，如果改变数据集，需要重新指定`Infer.PostProcess.class_id_map_file`，该映射文件的制作方法可以参考`ppcls/utils/imagenet1k_label_list.txt`。
|
||||
|
||||
|
||||
|
||||
<a name="3"></a>
|
||||
|
||||
## 3. 模型推理部署
|
||||
|
||||
<a name="3.1"></a>
|
||||
|
||||
### 3.1 推理模型准备
|
||||
|
||||
Paddle Inference 是飞桨的原生推理库， 作用于服务器端和云端，提供高性能的推理能力。相比于直接基于预训练模型进行预测，Paddle Inference可使用MKLDNN、CUDNN、TensorRT 进行预测加速，从而实现更优的推理性能。更多关于Paddle Inference推理引擎的介绍，可以参考[Paddle Inference官网教程](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html)。
|
||||
|
||||
|
||||
此处，我们提供了将权重和模型转换的脚本，执行该脚本可以得到对应的 inference 模型：
|
||||
|
||||
```shell
|
||||
python3 tools/export_model.py \
|
||||
-c ppcls/configs/ImageNet/PPHGNetV2/PPHGNetV2_B4.yaml \
|
||||
-o Global.pretrained_model=output/PPHGNetV2_B4/best_model \
|
||||
-o Global.save_inference_dir=deploy/models/PPHGNetV2_B4_infer
|
||||
```
|
||||
执行完该脚本后会在 `deploy/models/` 下生成 `PPHGNetV2_B4_infer` 文件夹，`models` 文件夹下应有如下文件结构：
|
||||
|
||||
```
|
||||
├── PPHGNetV2_B4_infer
|
||||
│ ├── inference.pdiparams
|
||||
│ ├── inference.pdiparams.info
|
||||
│ └── inference.pdmodel
|
||||
```
|
||||
|
||||
<a name="3.2"></a>
|
||||
|
||||
### 3.2 基于 Python 预测引擎推理
|
||||
|
||||
|
||||
<a name="3.2.1"></a>
|
||||
|
||||
#### 3.2.1 预测单张图像
|
||||
|
||||
返回 `deploy` 目录：
|
||||
|
||||
```
|
||||
cd ../
|
||||
```
|
||||
|
||||
运行下面的命令，对图像 `./images/ImageNet/ILSVRC2012_val_00000010.jpeg` 进行分类。
|
||||
|
||||
```shell
|
||||
# 使用下面的命令使用 GPU 进行预测
|
||||
python3 python/predict_cls.py -c configs/inference_cls.yaml -o Global.inference_model_dir=models/PPHGNetV2_B4_infer
|
||||
# 使用下面的命令使用 CPU 进行预测
|
||||
python3 python/predict_cls.py -c configs/inference_cls.yaml -o Global.inference_model_dir=models/PPHGNetV2_B4_infer -o Global.use_gpu=False
|
||||
```
|
||||
|
||||
输出结果如下。
|
||||
|
||||
```
|
||||
ILSVRC2012_val_00000010.jpeg: class id(s): [332, 153, 283, 338, 265], score(s): [0.94, 0.03, 0.02, 0.00, 0.00], label_name(s): ['Angora, Angora rabbit', 'Maltese dog, Maltese terrier, Maltese', 'Persian cat', 'guinea pig, Cavia cobaya', 'toy poodle']
|
||||
```
|
||||
|
||||
<a name="3.2.2"></a>
|
||||
|
||||
#### 3.2.2 基于文件夹的批量预测
|
||||
|
||||
如果希望预测文件夹内的图像，可以直接修改配置文件中的 `Global.infer_imgs` 字段，也可以通过下面的 `-o` 参数修改对应的配置。
|
||||
|
||||
```shell
|
||||
# 使用下面的命令使用 GPU 进行预测，如果希望使用 CPU 预测，可以在命令后面添加 -o Global.use_gpu=False
|
||||
python3 python/predict_cls.py -c configs/inference_cls.yaml -o Global.inference_model_dir=models/PPHGNetV2_B4_infer -o Global.infer_imgs=images/ImageNet/
|
||||
```
|
||||
|
||||
终端中会输出该文件夹内所有图像的分类结果，如下所示。
|
||||
|
||||
```
|
||||
ILSVRC2012_val_00000010.jpeg: class id(s): [332, 153, 283, 338, 265], score(s): [0.94, 0.03, 0.02, 0.00, 0.00], label_name(s): ['Angora, Angora rabbit', 'Maltese dog, Maltese terrier, Maltese', 'Persian cat', 'guinea pig, Cavia cobaya', 'toy poodle']
|
||||
ILSVRC2012_val_00010010.jpeg: class id(s): [626, 487, 531, 622, 593], score(s): [0.81, 0.08, 0.03, 0.01, 0.01], label_name(s): ['lighter, light, igniter, ignitor', 'cellular telephone, cellular phone, cellphone, cell, mobile phone', 'digital watch', 'lens cap, lens cover', 'harmonica, mouth organ, harp, mouth harp']
|
||||
ILSVRC2012_val_00020010.jpeg: class id(s): [178, 211, 246, 236, 181], score(s): [1.00, 0.00, 0.00, 0.00, 0.00], label_name(s): ['Weimaraner', 'vizsla, Hungarian pointer', 'Great Dane', 'Doberman, Doberman pinscher', 'Bedlington terrier']
|
||||
ILSVRC2012_val_00030010.jpeg: class id(s): [80, 83, 23, 8, 81], score(s): [1.00, 0.00, 0.00, 0.00, 0.00], label_name(s): ['black grouse', 'prairie chicken, prairie grouse, prairie fowl', 'vulture', 'hen', 'ptarmigan']
|
||||
```
|
||||
|
||||
<a name="3.3"></a>
|
||||
|
||||
### 3.3 基于 C++ 预测引擎推理
|
||||
|
||||
PaddleClas 提供了基于 C++ 预测引擎推理的示例，您可以参考[服务器端 C++ 预测](../../deployment/image_classification/cpp/linux.md)来完成相应的推理部署。如果您使用的是 Windows 平台，可以参考[基于 Visual Studio 2019 Community CMake 编译指南](../../deployment/image_classification/cpp/windows.md)完成相应的预测库编译和模型预测工作。
|
||||
|
||||
<a name="3.4"></a>
|
||||
|
||||
### 3.4 服务化部署
|
||||
|
||||
Paddle Serving 提供高性能、灵活易用的工业级在线推理服务。Paddle Serving 支持 RESTful、gRPC、bRPC 等多种协议，提供多种异构硬件和多种操作系统环境下推理解决方案。更多关于Paddle Serving 的介绍，可以参考[Paddle Serving 代码仓库](https://github.com/PaddlePaddle/Serving)。
|
||||
|
||||
PaddleClas 提供了基于 Paddle Serving 来完成模型服务化部署的示例，您可以参考[模型服务化部署](../../deployment/image_classification/paddle_serving.md)来完成相应的部署工作。
|
||||
|
||||
<a name="3.5"></a>
|
||||
|
||||
### 3.5 端侧部署
|
||||
|
||||
Paddle Lite 是一个高性能、轻量级、灵活性强且易于扩展的深度学习推理框架，定位于支持包括移动端、嵌入式以及服务器端在内的多硬件平台。更多关于 Paddle Lite 的介绍，可以参考[Paddle Lite 代码仓库](https://github.com/PaddlePaddle/Paddle-Lite)。
|
||||
|
||||
PaddleClas 提供了基于 Paddle Lite 来完成模型端侧部署的示例，您可以参考[端侧部署](../../deployment/image_classification/paddle_lite.md)来完成相应的部署工作。
|
||||
|
||||
<a name="3.6"></a>
|
||||
|
||||
### 3.6 Paddle2ONNX 模型转换与预测
|
||||
|
||||
Paddle2ONNX 支持将 PaddlePaddle 模型格式转化到 ONNX 模型格式。通过 ONNX 可以完成将 Paddle 模型到多种推理引擎的部署，包括TensorRT/OpenVINO/MNN/TNN/NCNN，以及其它对 ONNX 开源格式进行支持的推理引擎或硬件。更多关于 Paddle2ONNX 的介绍，可以参考[Paddle2ONNX 代码仓库](https://github.com/PaddlePaddle/Paddle2ONNX)。
|
||||
|
||||
PaddleClas 提供了基于 Paddle2ONNX 来完成 inference 模型转换 ONNX 模型并作推理预测的示例，您可以参考[Paddle2ONNX 模型转换与预测](../../deployment/image_classification/paddle2onnx.md)来完成相应的部署工作。
|
||||
|
|
@ -25,6 +25,7 @@ from .legendary_models.pp_lcnet import PPLCNet_x0_25, PPLCNet_x0_35, PPLCNet_x0_
|
|||
from .legendary_models.pp_lcnet_v2 import PPLCNetV2_small, PPLCNetV2_base, PPLCNetV2_large
|
||||
from .legendary_models.esnet import ESNet_x0_25, ESNet_x0_5, ESNet_x0_75, ESNet_x1_0
|
||||
from .legendary_models.pp_hgnet import PPHGNet_tiny, PPHGNet_small, PPHGNet_base
|
||||
from .legendary_models.pp_hgnet_v2 import PPHGNetV2_B0, PPHGNetV2_B1, PPHGNetV2_B2, PPHGNetV2_B3, PPHGNetV2_B4, PPHGNetV2_B5, PPHGNetV2_B6, PPHGNetV2_B7
|
||||
|
||||
from .model_zoo.resnet_vc import ResNet50_vc
|
||||
from .model_zoo.resnext import ResNeXt50_32x4d, ResNeXt50_64x4d, ResNeXt101_32x4d, ResNeXt101_64x4d, ResNeXt152_32x4d, ResNeXt152_64x4d
|
||||
|
|
|
|||
|
|
@ -0,0 +1,731 @@
|
|||
# copyright (c) 2023 PaddlePaddle Authors. All Rights Reserve.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
# You may obtain a copy of the License at
|
||||
#
|
||||
# http://www.apache.org/licenses/LICENSE-2.0
|
||||
#
|
||||
# Unless required by applicable law or agreed to in writing, software
|
||||
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
|
||||
import paddle
|
||||
import paddle.nn as nn
|
||||
import paddle.nn.functional as F
|
||||
from paddle.nn.initializer import KaimingNormal, Constant
|
||||
from paddle.nn import Conv2D, BatchNorm2D, ReLU, AdaptiveAvgPool2D, MaxPool2D
|
||||
from paddle.regularizer import L2Decay
|
||||
from paddle import ParamAttr
|
||||
|
||||
from ..base.theseus_layer import TheseusLayer
|
||||
from ....utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
|
||||
|
||||
MODEL_URLS = {
|
||||
"PPHGNetV2_B0":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B0_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B1":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B1_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B2":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B2_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B3":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B3_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B4":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B4_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B5":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B5_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B6":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B6_ssld_pretrained.pdparams",
|
||||
"PPHGNetV2_B7":
|
||||
"https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPHGNetV2_B7_ssld_pretrained.pdparams",
|
||||
}
|
||||
|
||||
__all__ = list(MODEL_URLS.keys())
|
||||
|
||||
kaiming_normal_ = KaimingNormal()
|
||||
zeros_ = Constant(value=0.)
|
||||
ones_ = Constant(value=1.)
|
||||
|
||||
|
||||
class LearnableAffineBlock(TheseusLayer):
|
||||
"""
|
||||
Create a learnable affine block module. This module can significantly improve accuracy on smaller models.
|
||||
|
||||
Args:
|
||||
scale_value (float): The initial value of the scale parameter, default is 1.0.
|
||||
bias_value (float): The initial value of the bias parameter, default is 0.0.
|
||||
lr_mult (float): The learning rate multiplier, default is 1.0.
|
||||
lab_lr (float): The learning rate, default is 0.01.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
scale_value=1.0,
|
||||
bias_value=0.0,
|
||||
lr_mult=1.0,
|
||||
lab_lr=0.01):
|
||||
super().__init__()
|
||||
self.scale = self.create_parameter(
|
||||
shape=[1, ],
|
||||
default_initializer=Constant(value=scale_value),
|
||||
attr=ParamAttr(learning_rate=lr_mult * lab_lr))
|
||||
self.add_parameter("scale", self.scale)
|
||||
self.bias = self.create_parameter(
|
||||
shape=[1, ],
|
||||
default_initializer=Constant(value=bias_value),
|
||||
attr=ParamAttr(learning_rate=lr_mult * lab_lr))
|
||||
self.add_parameter("bias", self.bias)
|
||||
|
||||
def forward(self, x):
|
||||
return self.scale * x + self.bias
|
||||
|
||||
|
||||
class ConvBNAct(TheseusLayer):
|
||||
"""
|
||||
ConvBNAct is a combination of convolution and batchnorm layers.
|
||||
|
||||
Args:
|
||||
in_channels (int): Number of input channels.
|
||||
out_channels (int): Number of output channels.
|
||||
kernel_size (int): Size of the convolution kernel. Defaults to 3.
|
||||
stride (int): Stride of the convolution. Defaults to 1.
|
||||
padding (int/str): Padding or padding type for the convolution. Defaults to 1.
|
||||
groups (int): Number of groups for the convolution. Defaults to 1.
|
||||
use_act: (bool): Whether to use activation function. Defaults to True.
|
||||
use_lab (bool): Whether to use the LAB operation. Defaults to False.
|
||||
lr_mult (float): Learning rate multiplier for the layer. Defaults to 1.0.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
in_channels,
|
||||
out_channels,
|
||||
kernel_size=3,
|
||||
stride=1,
|
||||
padding=1,
|
||||
groups=1,
|
||||
use_act=True,
|
||||
use_lab=False,
|
||||
lr_mult=1.0):
|
||||
super().__init__()
|
||||
self.use_act = use_act
|
||||
self.use_lab = use_lab
|
||||
self.conv = Conv2D(
|
||||
in_channels,
|
||||
out_channels,
|
||||
kernel_size,
|
||||
stride,
|
||||
padding=padding
|
||||
if isinstance(padding, str) else (kernel_size - 1) // 2,
|
||||
groups=groups,
|
||||
weight_attr=ParamAttr(learning_rate=lr_mult),
|
||||
bias_attr=False)
|
||||
self.bn = BatchNorm2D(
|
||||
out_channels,
|
||||
weight_attr=ParamAttr(
|
||||
regularizer=L2Decay(0.0), learning_rate=lr_mult),
|
||||
bias_attr=ParamAttr(
|
||||
regularizer=L2Decay(0.0), learning_rate=lr_mult))
|
||||
if self.use_act:
|
||||
self.act = ReLU()
|
||||
if self.use_lab:
|
||||
self.lab = LearnableAffineBlock(lr_mult=lr_mult)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.conv(x)
|
||||
x = self.bn(x)
|
||||
if self.use_act:
|
||||
x = self.act(x)
|
||||
if self.use_lab:
|
||||
x = self.lab(x)
|
||||
return x
|
||||
|
||||
|
||||
class LightConvBNAct(TheseusLayer):
|
||||
"""
|
||||
LightConvBNAct is a combination of pw and dw layers.
|
||||
|
||||
Args:
|
||||
in_channels (int): Number of input channels.
|
||||
out_channels (int): Number of output channels.
|
||||
kernel_size (int): Size of the depth-wise convolution kernel.
|
||||
use_lab (bool): Whether to use the LAB operation. Defaults to False.
|
||||
lr_mult (float): Learning rate multiplier for the layer. Defaults to 1.0.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
in_channels,
|
||||
out_channels,
|
||||
kernel_size,
|
||||
use_lab=False,
|
||||
lr_mult=1.0,
|
||||
**kwargs):
|
||||
super().__init__()
|
||||
self.conv1 = ConvBNAct(
|
||||
in_channels=in_channels,
|
||||
out_channels=out_channels,
|
||||
kernel_size=1,
|
||||
use_act=False,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.conv2 = ConvBNAct(
|
||||
in_channels=out_channels,
|
||||
out_channels=out_channels,
|
||||
kernel_size=kernel_size,
|
||||
groups=out_channels,
|
||||
use_act=True,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.conv1(x)
|
||||
x = self.conv2(x)
|
||||
return x
|
||||
|
||||
|
||||
class StemBlock(TheseusLayer):
|
||||
"""
|
||||
StemBlock for PP-HGNetV2.
|
||||
|
||||
Args:
|
||||
in_channels (int): Number of input channels.
|
||||
mid_channels (int): Number of middle channels.
|
||||
out_channels (int): Number of output channels.
|
||||
use_lab (bool): Whether to use the LAB operation. Defaults to False.
|
||||
lr_mult (float): Learning rate multiplier for the layer. Defaults to 1.0.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
in_channels,
|
||||
mid_channels,
|
||||
out_channels,
|
||||
use_lab=False,
|
||||
lr_mult=1.0):
|
||||
super().__init__()
|
||||
self.stem1 = ConvBNAct(
|
||||
in_channels=in_channels,
|
||||
out_channels=mid_channels,
|
||||
kernel_size=3,
|
||||
stride=2,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.stem2a = ConvBNAct(
|
||||
in_channels=mid_channels,
|
||||
out_channels=mid_channels // 2,
|
||||
kernel_size=2,
|
||||
stride=1,
|
||||
padding="SAME",
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.stem2b = ConvBNAct(
|
||||
in_channels=mid_channels // 2,
|
||||
out_channels=mid_channels,
|
||||
kernel_size=2,
|
||||
stride=1,
|
||||
padding="SAME",
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.stem3 = ConvBNAct(
|
||||
in_channels=mid_channels * 2,
|
||||
out_channels=mid_channels,
|
||||
kernel_size=3,
|
||||
stride=2,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.stem4 = ConvBNAct(
|
||||
in_channels=mid_channels,
|
||||
out_channels=out_channels,
|
||||
kernel_size=1,
|
||||
stride=1,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.pool = nn.MaxPool2D(
|
||||
kernel_size=2, stride=1, ceil_mode=True, padding="SAME")
|
||||
|
||||
def forward(self, x):
|
||||
x = self.stem1(x)
|
||||
x2 = self.stem2a(x)
|
||||
x2 = self.stem2b(x2)
|
||||
x1 = self.pool(x)
|
||||
x = paddle.concat([x1, x2], 1)
|
||||
x = self.stem3(x)
|
||||
x = self.stem4(x)
|
||||
|
||||
return x
|
||||
|
||||
|
||||
class HGV2_Block(TheseusLayer):
|
||||
"""
|
||||
HGV2_Block, the basic unit that constitutes the HGV2_Stage.
|
||||
|
||||
Args:
|
||||
in_channels (int): Number of input channels.
|
||||
mid_channels (int): Number of middle channels.
|
||||
out_channels (int): Number of output channels.
|
||||
kernel_size (int): Size of the convolution kernel. Defaults to 3.
|
||||
layer_num (int): Number of layers in the HGV2 block. Defaults to 6.
|
||||
stride (int): Stride of the convolution. Defaults to 1.
|
||||
padding (int/str): Padding or padding type for the convolution. Defaults to 1.
|
||||
groups (int): Number of groups for the convolution. Defaults to 1.
|
||||
use_act (bool): Whether to use activation function. Defaults to True.
|
||||
use_lab (bool): Whether to use the LAB operation. Defaults to False.
|
||||
lr_mult (float): Learning rate multiplier for the layer. Defaults to 1.0.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
in_channels,
|
||||
mid_channels,
|
||||
out_channels,
|
||||
kernel_size=3,
|
||||
layer_num=6,
|
||||
identity=False,
|
||||
light_block=True,
|
||||
use_lab=False,
|
||||
lr_mult=1.0):
|
||||
super().__init__()
|
||||
self.identity = identity
|
||||
|
||||
self.layers = nn.LayerList()
|
||||
block_type = "LightConvBNAct" if light_block else "ConvBNAct"
|
||||
for i in range(layer_num):
|
||||
self.layers.append(
|
||||
eval(block_type)(in_channels=in_channels
|
||||
if i == 0 else mid_channels,
|
||||
out_channels=mid_channels,
|
||||
stride=1,
|
||||
kernel_size=kernel_size,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult))
|
||||
# feature aggregation
|
||||
total_channels = in_channels + layer_num * mid_channels
|
||||
self.aggregation_squeeze_conv = ConvBNAct(
|
||||
in_channels=total_channels,
|
||||
out_channels=out_channels // 2,
|
||||
kernel_size=1,
|
||||
stride=1,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
self.aggregation_excitation_conv = ConvBNAct(
|
||||
in_channels=out_channels // 2,
|
||||
out_channels=out_channels,
|
||||
kernel_size=1,
|
||||
stride=1,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
|
||||
def forward(self, x):
|
||||
identity = x
|
||||
output = []
|
||||
output.append(x)
|
||||
for layer in self.layers:
|
||||
x = layer(x)
|
||||
output.append(x)
|
||||
x = paddle.concat(output, axis=1)
|
||||
x = self.aggregation_squeeze_conv(x)
|
||||
x = self.aggregation_excitation_conv(x)
|
||||
if self.identity:
|
||||
x += identity
|
||||
return x
|
||||
|
||||
|
||||
class HGV2_Stage(TheseusLayer):
|
||||
"""
|
||||
HGV2_Stage, the basic unit that constitutes the PPHGNetV2.
|
||||
|
||||
Args:
|
||||
in_channels (int): Number of input channels.
|
||||
mid_channels (int): Number of middle channels.
|
||||
out_channels (int): Number of output channels.
|
||||
block_num (int): Number of blocks in the HGV2 stage.
|
||||
layer_num (int): Number of layers in the HGV2 block. Defaults to 6.
|
||||
is_downsample (bool): Whether to use downsampling operation. Defaults to False.
|
||||
light_block (bool): Whether to use light block. Defaults to True.
|
||||
kernel_size (int): Size of the convolution kernel. Defaults to 3.
|
||||
use_lab (bool, optional): Whether to use the LAB operation. Defaults to False.
|
||||
lr_mult (float, optional): Learning rate multiplier for the layer. Defaults to 1.0.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
in_channels,
|
||||
mid_channels,
|
||||
out_channels,
|
||||
block_num,
|
||||
layer_num=6,
|
||||
is_downsample=True,
|
||||
light_block=True,
|
||||
kernel_size=3,
|
||||
use_lab=False,
|
||||
lr_mult=1.0):
|
||||
|
||||
super().__init__()
|
||||
self.is_downsample = is_downsample
|
||||
if self.is_downsample:
|
||||
self.downsample = ConvBNAct(
|
||||
in_channels=in_channels,
|
||||
out_channels=in_channels,
|
||||
kernel_size=3,
|
||||
stride=2,
|
||||
groups=in_channels,
|
||||
use_act=False,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult)
|
||||
|
||||
blocks_list = []
|
||||
for i in range(block_num):
|
||||
blocks_list.append(
|
||||
HGV2_Block(
|
||||
in_channels=in_channels if i == 0 else out_channels,
|
||||
mid_channels=mid_channels,
|
||||
out_channels=out_channels,
|
||||
kernel_size=kernel_size,
|
||||
layer_num=layer_num,
|
||||
identity=False if i == 0 else True,
|
||||
light_block=light_block,
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult))
|
||||
self.blocks = nn.Sequential(*blocks_list)
|
||||
|
||||
def forward(self, x):
|
||||
if self.is_downsample:
|
||||
x = self.downsample(x)
|
||||
x = self.blocks(x)
|
||||
return x
|
||||
|
||||
|
||||
class PPHGNetV2(TheseusLayer):
|
||||
"""
|
||||
PPHGNetV2
|
||||
|
||||
Args:
|
||||
stage_config (dict): Config for PPHGNetV2 stages. such as the number of channels, stride, etc.
|
||||
stem_channels: (list): Number of channels of the stem of the PPHGNetV2.
|
||||
use_lab (bool): Whether to use the LAB operation. Defaults to False.
|
||||
use_last_conv (bool): Whether to use the last conv layer as the output channel. Defaults to True.
|
||||
class_expand (int): Number of channels for the last 1x1 convolutional layer.
|
||||
drop_prob (float): Dropout probability for the last 1x1 convolutional layer. Defaults to 0.0.
|
||||
class_num (int): The number of classes for the classification layer. Defaults to 1000.
|
||||
lr_mult_list (list): Learning rate multiplier for the stages. Defaults to [1.0, 1.0, 1.0, 1.0, 1.0].
|
||||
Returns:
|
||||
model: nn.Layer. Specific PPHGNetV2 model depends on args.
|
||||
"""
|
||||
|
||||
def __init__(self,
|
||||
stage_config,
|
||||
stem_channels=[3, 32, 64],
|
||||
use_lab=False,
|
||||
use_last_conv=True,
|
||||
class_expand=2048,
|
||||
dropout_prob=0.0,
|
||||
class_num=1000,
|
||||
lr_mult_list=[1.0, 1.0, 1.0, 1.0, 1.0],
|
||||
**kwargs):
|
||||
super().__init__()
|
||||
self.use_lab = use_lab
|
||||
self.use_last_conv = use_last_conv
|
||||
self.class_expand = class_expand
|
||||
self.class_num = class_num
|
||||
|
||||
# stem
|
||||
self.stem = StemBlock(
|
||||
in_channels=stem_channels[0],
|
||||
mid_channels=stem_channels[1],
|
||||
out_channels=stem_channels[2],
|
||||
use_lab=use_lab,
|
||||
lr_mult=lr_mult_list[0])
|
||||
|
||||
# stages
|
||||
self.stages = nn.LayerList()
|
||||
for i, k in enumerate(stage_config):
|
||||
in_channels, mid_channels, out_channels, block_num, is_downsample, light_block, kernel_size, layer_num = stage_config[
|
||||
k]
|
||||
self.stages.append(
|
||||
HGV2_Stage(
|
||||
in_channels,
|
||||
mid_channels,
|
||||
out_channels,
|
||||
block_num,
|
||||
layer_num,
|
||||
is_downsample,
|
||||
light_block,
|
||||
kernel_size,
|
||||
use_lab,
|
||||
lr_mult=lr_mult_list[i + 1]))
|
||||
|
||||
self.avg_pool = AdaptiveAvgPool2D(1)
|
||||
|
||||
if self.use_last_conv:
|
||||
self.last_conv = Conv2D(
|
||||
in_channels=out_channels,
|
||||
out_channels=self.class_expand,
|
||||
kernel_size=1,
|
||||
stride=1,
|
||||
padding=0,
|
||||
bias_attr=False)
|
||||
self.act = ReLU()
|
||||
if self.use_lab:
|
||||
self.lab = LearnableAffineBlock()
|
||||
self.dropout = nn.Dropout(
|
||||
p=dropout_prob, mode="downscale_in_infer")
|
||||
|
||||
self.flatten = nn.Flatten(start_axis=1, stop_axis=-1)
|
||||
self.fc = nn.Linear(self.class_expand if self.use_last_conv else
|
||||
out_channels, self.class_num)
|
||||
|
||||
self._init_weights()
|
||||
|
||||
def _init_weights(self):
|
||||
for m in self.sublayers():
|
||||
if isinstance(m, nn.Conv2D):
|
||||
kaiming_normal_(m.weight)
|
||||
elif isinstance(m, (nn.BatchNorm2D)):
|
||||
ones_(m.weight)
|
||||
zeros_(m.bias)
|
||||
elif isinstance(m, nn.Linear):
|
||||
zeros_(m.bias)
|
||||
|
||||
def forward(self, x):
|
||||
x = self.stem(x)
|
||||
|
||||
for stage in self.stages:
|
||||
x = stage(x)
|
||||
x = self.avg_pool(x)
|
||||
|
||||
if self.use_last_conv:
|
||||
x = self.last_conv(x)
|
||||
x = self.act(x)
|
||||
if self.use_lab:
|
||||
x = self.lab(x)
|
||||
x = self.dropout(x)
|
||||
x = self.flatten(x)
|
||||
x = self.fc(x)
|
||||
|
||||
return x
|
||||
|
||||
|
||||
def _load_pretrained(pretrained, model, model_url, use_ssld):
|
||||
if pretrained is False:
|
||||
pass
|
||||
elif pretrained is True:
|
||||
load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
|
||||
elif isinstance(pretrained, str):
|
||||
load_dygraph_pretrain(model, pretrained)
|
||||
else:
|
||||
raise RuntimeError(
|
||||
"pretrained type is not available. Please use `string` or `boolean` type."
|
||||
)
|
||||
|
||||
|
||||
def PPHGNetV2_B0(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B0
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B0` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [16, 16, 64, 1, False, False, 3, 3],
|
||||
"stage2": [64, 32, 256, 1, True, False, 3, 3],
|
||||
"stage3": [256, 64, 512, 2, True, True, 5, 3],
|
||||
"stage4": [512, 128, 1024, 1, True, True, 5, 3],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 16, 16],
|
||||
stage_config=stage_config,
|
||||
use_lab=True,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B0"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B1(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B1
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B1` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [32, 32, 64, 1, False, False, 3, 3],
|
||||
"stage2": [64, 48, 256, 1, True, False, 3, 3],
|
||||
"stage3": [256, 96, 512, 2, True, True, 5, 3],
|
||||
"stage4": [512, 192, 1024, 1, True, True, 5, 3],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 24, 32],
|
||||
stage_config=stage_config,
|
||||
use_lab=True,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B1"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B2(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B2
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B2` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [32, 32, 96, 1, False, False, 3, 4],
|
||||
"stage2": [96, 64, 384, 1, True, False, 3, 4],
|
||||
"stage3": [384, 128, 768, 3, True, True, 5, 4],
|
||||
"stage4": [768, 256, 1536, 1, True, True, 5, 4],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 24, 32],
|
||||
stage_config=stage_config,
|
||||
use_lab=True,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B2"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B3(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B3
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B3` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [32, 32, 128, 1, False, False, 3, 5],
|
||||
"stage2": [128, 64, 512, 1, True, False, 3, 5],
|
||||
"stage3": [512, 128, 1024, 3, True, True, 5, 5],
|
||||
"stage4": [1024, 256, 2048, 1, True, True, 5, 5],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 24, 32],
|
||||
stage_config=stage_config,
|
||||
use_lab=True,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B3"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B4(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B4
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B4` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [48, 48, 128, 1, False, False, 3, 6],
|
||||
"stage2": [128, 96, 512, 1, True, False, 3, 6],
|
||||
"stage3": [512, 192, 1024, 3, True, True, 5, 6],
|
||||
"stage4": [1024, 384, 2048, 1, True, True, 5, 6],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 32, 48],
|
||||
stage_config=stage_config,
|
||||
use_lab=False,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B4"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B5(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B5
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B5` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [64, 64, 128, 1, False, False, 3, 6],
|
||||
"stage2": [128, 128, 512, 2, True, False, 3, 6],
|
||||
"stage3": [512, 256, 1024, 5, True, True, 5, 6],
|
||||
"stage4": [1024, 512, 2048, 2, True, True, 5, 6],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 32, 64],
|
||||
stage_config=stage_config,
|
||||
use_lab=False,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B5"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B6(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B6
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B6` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [96, 96, 192, 2, False, False, 3, 6],
|
||||
"stage2": [192, 192, 512, 3, True, False, 3, 6],
|
||||
"stage3": [512, 384, 1024, 6, True, True, 5, 6],
|
||||
"stage4": [1024, 768, 2048, 3, True, True, 5, 6],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 48, 96],
|
||||
stage_config=stage_config,
|
||||
use_lab=False,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B6"], use_ssld)
|
||||
return model
|
||||
|
||||
|
||||
def PPHGNetV2_B7(pretrained=False, use_ssld=False, **kwargs):
|
||||
"""
|
||||
PPHGNetV2_B7
|
||||
Args:
|
||||
pretrained (bool/str): If `True` load pretrained parameters, `False` otherwise.
|
||||
If str, means the path of the pretrained model.
|
||||
use_ssld (bool) Whether using ssld pretrained model when pretrained is True.
|
||||
Returns:
|
||||
model: nn.Layer. Specific `PPHGNetV2_B7` model depends on args.
|
||||
"""
|
||||
stage_config = {
|
||||
# in_channels, mid_channels, out_channels, num_blocks, is_downsample, light_block, kernel_size, layer_num
|
||||
"stage1": [128, 128, 256, 2, False, False, 3, 7],
|
||||
"stage2": [256, 256, 512, 4, True, False, 3, 7],
|
||||
"stage3": [512, 512, 1024, 12, True, True, 5, 7],
|
||||
"stage4": [1024, 1024, 2048, 4, True, True, 5, 7],
|
||||
}
|
||||
|
||||
model = PPHGNetV2(
|
||||
stem_channels=[3, 64, 128],
|
||||
stage_config=stage_config,
|
||||
use_lab=False,
|
||||
**kwargs)
|
||||
_load_pretrained(pretrained, model, MODEL_URLS["PPHGNetV2_B7"], use_ssld)
|
||||
return model
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B0
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B1
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B2
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B3
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B4
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,172 @@
|
|||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: "./output/"
|
||||
device: "gpu"
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 360
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: "./inference"
|
||||
use_dali: false
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: "DistillationModel"
|
||||
class_num: &class_num 1000
|
||||
# if not null, its lengths should be same as models
|
||||
pretrained_list:
|
||||
# if not null, its lengths should be same as models
|
||||
freeze_params_list:
|
||||
- True
|
||||
- False
|
||||
models:
|
||||
- Teacher:
|
||||
name: PPHGNet_small
|
||||
class_num: *class_num
|
||||
pretrained: True
|
||||
use_ssld: True
|
||||
- Student:
|
||||
name: PPHGNetV2_B4
|
||||
class_num: *class_num
|
||||
pretrained: False
|
||||
|
||||
infer_model_name: "Student"
|
||||
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- DistillationCELoss:
|
||||
weight: 1.0
|
||||
model_name_pairs:
|
||||
- ["Student", "Teacher"]
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: "./dataset/"
|
||||
# ImageNet_5M label path, the training process does not use real labels.
|
||||
cls_label_path: "./dataset/train_list_imagenet_5M.txt"
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 0.00392157
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 8
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: "./dataset/ILSVRC2012/"
|
||||
cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 236
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 0.00392157
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 8
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: "docs/images/inference_deployment/whl_demo.jpg"
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 236
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: "ppcls/utils/imagenet1k_label_list.txt"
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- DistillationTopkAcc:
|
||||
model_key: "Student"
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- DistillationTopkAcc:
|
||||
model_key: "Student"
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,173 @@
|
|||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: "./output/"
|
||||
device: "gpu"
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 60
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: "./inference"
|
||||
use_dali: false
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: "DistillationModel"
|
||||
class_num: &class_num 1000
|
||||
# if not null, its lengths should be same as models
|
||||
pretrained_list:
|
||||
# if not null, its lengths should be same as models
|
||||
freeze_params_list:
|
||||
- True
|
||||
- False
|
||||
models:
|
||||
- Teacher:
|
||||
name: PPHGNet_small
|
||||
class_num: *class_num
|
||||
pretrained: True
|
||||
use_ssld: True
|
||||
- Student:
|
||||
name: PPHGNetV2_B4
|
||||
class_num: *class_num
|
||||
pretrained: path/to/stage1_best_model_student
|
||||
|
||||
infer_model_name: "Student"
|
||||
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- DistillationCELoss:
|
||||
weight: 1.0
|
||||
model_name_pairs:
|
||||
- ["Student", "Teacher"]
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# stage2 should reduce learning rate
|
||||
learning_rate: 0.005
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: "./dataset/ILSVRC2012/"
|
||||
# ImageNet-1k label path, the training process does not use real labels
|
||||
cls_label_path: "./dataset/ILSVRC2012/train_list.txt"
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 0.00392157
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 8
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: "./dataset/ILSVRC2012/"
|
||||
cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 236
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 0.00392157
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 8
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: "docs/images/inference_deployment/whl_demo.jpg"
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 236
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: "ppcls/utils/imagenet1k_label_list.txt"
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- DistillationTopkAcc:
|
||||
model_key: "Student"
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- DistillationTopkAcc:
|
||||
model_key: "Student"
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B5
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B6
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
|
|
@ -0,0 +1,164 @@
|
|||
## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
|
||||
# global configs
|
||||
Global:
|
||||
checkpoints: null
|
||||
pretrained_model: null
|
||||
output_dir: ./output/
|
||||
device: gpu
|
||||
save_interval: 1
|
||||
eval_during_train: True
|
||||
eval_interval: 1
|
||||
epochs: 200
|
||||
print_batch_step: 10
|
||||
use_visualdl: False
|
||||
# used for static mode and model export
|
||||
image_shape: [3, 224, 224]
|
||||
save_inference_dir: ./inference
|
||||
# training model under @to_static
|
||||
to_static: False
|
||||
use_dali: False
|
||||
|
||||
# mixed precision training
|
||||
AMP:
|
||||
use_amp: True
|
||||
use_fp16_test: False
|
||||
scale_loss: 128.0
|
||||
use_dynamic_loss_scaling: True
|
||||
use_promote: False
|
||||
# O1: mixed fp16, O2: pure fp16
|
||||
level: O1
|
||||
|
||||
# model architecture
|
||||
Arch:
|
||||
name: PPHGNetV2_B7
|
||||
class_num: 1000
|
||||
pretrained: True # ssld pretrained
|
||||
|
||||
# loss function config for traing/eval process
|
||||
Loss:
|
||||
Train:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
epsilon: 0.1
|
||||
Eval:
|
||||
- CELoss:
|
||||
weight: 1.0
|
||||
|
||||
|
||||
Optimizer:
|
||||
name: Momentum
|
||||
momentum: 0.9
|
||||
lr:
|
||||
name: Cosine
|
||||
# for global bs 1024, when finetune training, you need to reduce learning_rate manually
|
||||
learning_rate: 0.5
|
||||
warmup_epoch: 5
|
||||
regularizer:
|
||||
name: 'L2'
|
||||
coeff: 0.00002
|
||||
|
||||
|
||||
# data loader for train and eval
|
||||
DataLoader:
|
||||
Train:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/train_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- RandCropImage:
|
||||
size: 224
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- RandFlipImage:
|
||||
flip_code: 1
|
||||
- TimmAutoAugment:
|
||||
config_str: rand-m7-mstd0.5-inc1
|
||||
interpolation: bicubic
|
||||
img_size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- RandomErasing:
|
||||
EPSILON: 0.25
|
||||
sl: 0.02
|
||||
sh: 1.0/3.0
|
||||
r1: 0.3
|
||||
attempt: 10
|
||||
use_log_aspect: True
|
||||
mode: pixel
|
||||
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: True
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Eval:
|
||||
dataset:
|
||||
name: ImageNetDataset
|
||||
image_root: ./dataset/ILSVRC2012/
|
||||
cls_label_path: ./dataset/ILSVRC2012/val_list.txt
|
||||
transform_ops:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
sampler:
|
||||
name: DistributedBatchSampler
|
||||
batch_size: 128
|
||||
drop_last: False
|
||||
shuffle: False
|
||||
loader:
|
||||
num_workers: 16
|
||||
use_shared_memory: True
|
||||
|
||||
Infer:
|
||||
infer_imgs: docs/images/inference_deployment/whl_demo.jpg
|
||||
batch_size: 10
|
||||
transforms:
|
||||
- DecodeImage:
|
||||
to_rgb: True
|
||||
channel_first: False
|
||||
- ResizeImage:
|
||||
resize_short: 232
|
||||
interpolation: bicubic
|
||||
backend: pil
|
||||
- CropImage:
|
||||
size: 224
|
||||
- NormalizeImage:
|
||||
scale: 1.0/255.0
|
||||
mean: [0.485, 0.456, 0.406]
|
||||
std: [0.229, 0.224, 0.225]
|
||||
order: ''
|
||||
- ToCHWImage:
|
||||
PostProcess:
|
||||
name: Topk
|
||||
topk: 5
|
||||
class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
|
||||
|
||||
Metric:
|
||||
Train:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Eval:
|
||||
- TopkAcc:
|
||||
topk: [1, 5]
|
||||
Loading…
Reference in New Issue