[Feat] Support saving and loading models in different formats

README.md

@@ -1 +1 @@
-README_CN.md
+README_CN.md

README_CN.md

@@ -11,7 +11,7 @@
 
 [![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
 [![Version](https://img.shields.io/github/release/PaddlePaddle/PaddleSeg.svg)](https://github.com/PaddlePaddle/PaddleSeg/releases)
-![python version](https://img.shields.io/badge/python-3.6+-orange.svg)
+![python version](https://img.shields.io/badge/python-3.8+-orange.svg)
 ![support os](https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-yellow.svg)
 ![stars](https://img.shields.io/github/stars/PaddlePaddle/PaddleSeg?color=ccf)
 </div>

README_EN.md

@@ -10,7 +10,7 @@ English | [简体中文](README_CN.md)
 
 [![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
 [![Version](https://img.shields.io/github/release/PaddlePaddle/PaddleSeg.svg)](https://github.com/PaddlePaddle/PaddleSeg/releases)
-![python version](https://img.shields.io/badge/python-3.6+-orange.svg)
+![python version](https://img.shields.io/badge/python-3.8+-orange.svg)
 ![support os](https://img.shields.io/badge/os-linux%2C%20win%2C%20mac-yellow.svg)
 ![stars](https://img.shields.io/github/stars/PaddlePaddle/PaddleSeg?color=ccf)
 

deploy/python/collect_dynamic_shape.py

@@ -62,7 +62,7 @@ def collect_dynamic_shape(args):
 
     # prepare config
     cfg = DeployConfig(args.config)
-    pred_cfg = PredictConfig(cfg.model, cfg.params)
+    pred_cfg = PredictConfig(cfg.model_dir, cfg.model_prefix)
     pred_cfg.enable_use_gpu(1000, 0)
     pred_cfg.collect_shape_range_info(args.dynamic_shape_path)
 

deploy/python/infer.py

@@ -52,7 +52,7 @@ def auto_tune(args, imgs, img_nums):
     num = min(len(imgs), img_nums)
 
     cfg = DeployConfig(args.cfg)
-    pred_cfg = PredictConfig(cfg.model, cfg.params)
+    pred_cfg = PredictConfig(cfg.model_dir, cfg.model_prefix)
     pred_cfg.enable_use_gpu(100, 0)
     if not args.print_detail:
         pred_cfg.disable_glog_info()
@@ -139,7 +139,7 @@ def __init__(self, args):
                 logger=logger)
 
     def _init_base_config(self):
-        self.pred_cfg = PredictConfig(self.cfg.model, self.cfg.params)
+        self.pred_cfg = PredictConfig(self.cfg.model_dir, self.cfg.model_prefix)
         if not self.args.print_detail:
             self.pred_cfg.disable_glog_info()
         self.pred_cfg.enable_memory_optim()

deploy/python/infer_dataset.py

@@ -191,7 +191,7 @@ def auto_tune(args, dataset, img_nums):
     num = min(len(dataset), img_nums)
 
     cfg = DeployConfig(args.cfg)
-    pred_cfg = PredictConfig(cfg.model, cfg.params)
+    pred_cfg = PredictConfig(cfg.model_dir, cfg.model_prefix)
     pred_cfg.enable_use_gpu(100, 0)
     if not args.print_detail:
         pred_cfg.disable_glog_info()

docs/install.md

@@ -8,20 +8,20 @@ English | [简体中文](install_cn.md)
 - pip/pip3(9.0.1+)，64-bit version
 - CUDA >= 10.2
 - cuDNN >= 7.6
-- PaddlePaddle (the version >= 2.4)
+- PaddlePaddle (the version >= 3.0.0b0)
 
 ## 2 Installation
 
 ### 2.1 Install PaddlePaddle
 
-Please refer to the [installation doc](https://www.paddlepaddle.org.cn/documentation/docs/en/install/index_en.html) to install PaddlePaddle (the version >= 2.4).
+Please refer to the [installation doc](https://www.paddlepaddle.org.cn/documentation/docs/en/install/index_en.html) to install PaddlePaddle (the version >= 3.0.0b0).
 
 Highly recommend you install the GPU version of PaddlePaddle, due to the large overhead of segmentation models, otherwise, it could be out of memory while running the models.
 
 For example, run the following command to install Paddle with pip for Linux, CUDA 10.2.
 
 ```
-python -m pip install paddlepaddle-gpu==2.4.2 -i https://pypi.tuna.tsinghua.edu.cn/simple
+python -m pip install paddlepaddle-gpu==3.0.0b0 -i https://pypi.tuna.tsinghua.edu.cn/simple
 ```
 
 

docs/install_cn.md

@@ -9,18 +9,18 @@
 - pip/pip3(9.0.1+)，64位版本
 - CUDA >= 10.2
 - cuDNN >= 7.6
-- PaddlePaddle (版本==2.5)
+- PaddlePaddle (版本>=3.0.0b0)
 
 ## 2 本地安装说明
 
 ### 2.1 安装PaddlePaddle
 
-请参考[快速安装文档](https://www.paddlepaddle.org.cn/install/quick)或者[详细安装文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/index_cn.html)，安装PaddlePaddle （要求不低于2.4版本，推荐安装最新版本）。
+请参考[快速安装文档](https://www.paddlepaddle.org.cn/install/quick)或者[详细安装文档](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/index_cn.html)，安装PaddlePaddle （要求不低于3.0.0b0版本，推荐安装最新版本）。
 
 比如Linux、CUDA 10.2，使用pip安装GPU版本，执行如下命令。
 
 ```
-python -m pip install paddlepaddle-gpu==2.4.2 -i https://pypi.tuna.tsinghua.edu.cn/simple
+python -m pip install paddlepaddle-gpu==3.0.0b0 -i https://pypi.tuna.tsinghua.edu.cn/simple
 ```
 
 使用如下命令验证PaddlePaddle是否安装成功，并且查看版本。

paddleseg/deploy/infer.py

@@ -34,12 +34,12 @@ def transforms(self):
         return self._transforms
 
     @property
-    def model(self):
-        return os.path.join(self._dir, self.dic['Deploy']['model'])
+    def model_dir(self):
+        return self._dir
 
     @property
-    def params(self):
-        return os.path.join(self._dir, self.dic['Deploy']['params'])
+    def model_prefix(self):
+        return self.dic['Deploy']['model_prefix']
 
     @staticmethod
     def load_transforms(t_list):

paddleseg/models/backbones/efficientformerv2.py

@@ -203,7 +203,7 @@ def forward_tokens(self, x):
     def forward(self, x):
         x = self.patch_embed(x)
         x = self.forward_tokens(x)
-        if self.mode is not 'multi_scale':
+        if self.mode != 'multi_scale':
             x = [
                 paddle.concat(
                     [

tools/export.py

@@ -97,8 +97,7 @@ def main(args):
     # TODO add test config
     deploy_info = {
         'Deploy': {
-            'model': save_name + '.pdmodel',
-            'params': save_name + '.pdiparams',
+            'model_prefix': save_name,
             'transforms': transforms,
             'input_shape': shape,
             'output_op': args.output_op,