Merge pull request #1305 from PaddlePaddle/develop

Merge develop into release/2.3

deploy/python/preprocess.py

@@ -79,7 +79,7 @@ class UnifiedResize(object):
             if isinstance(interpolation, str):
                 interpolation = _cv2_interp_from_str[interpolation.lower()]
             # compatible with opencv < version 4.4.0
-            elif not interpolation:
+            elif interpolation is None:
                 interpolation = cv2.INTER_LINEAR
             self.resize_func = partial(cv2.resize, interpolation=interpolation)
         elif backend.lower() == "pil":

docs/zh_CN/tutorials/getting_started_retrieval.md

@@ -214,7 +214,7 @@ python3 tools/export_model.py \
 
 其中，`Global.pretrained_model`用于指定模型文件路径，该路径仍无需包含模型文件后缀名（如[2.2 模型恢复训练](#模型恢复训练)）。当执行后，会在当前目录下生成`./inference`目录，目录下包含`inference.pdiparams`、`inference.pdiparams.info`、`inference.pdmodel`文件。`Global.save_inference_dir`可以指定导出inference模型的路径。此处保存的inference模型在embedding特征层做了截断，即模型最终的输出为n维embedding特征。
 
-上述命令将生成模型结构文件（`inference.pdmodel`）和模型权重文件（`inference.pdiparams`），然后可以使用预测引擎进行推理。使用inference模型推理的流程可以参考[基于Python预测引擎预测推理](@shengyu)。
+上述命令将生成模型结构文件（`inference.pdmodel`）和模型权重文件（`inference.pdiparams`），然后可以使用预测引擎进行推理。使用inference模型推理的流程可以参考[基于Python预测引擎预测推理](https://github.com/PaddlePaddle/PaddleClas/blob/develop/docs/zh_CN/inference.md)。
 
 ## 基础知识
 

docs/zh_CN_tmp/image_recognition_pipeline/mainbody_detection.md

+# 主体检测
+
+
+主体检测技术是目前应用非常广泛的一种检测技术，它指的是检测出图片中一个或者多个主体的坐标位置，然后将图像中的对应区域裁剪下来，进行识别，从而完成整个识别过程。主体检测是识别任务的前序步骤，可以有效提升识别精度。
+
+本部分主要从数据集、模型选择和模型训练 3 个方面对该部分内容进行介绍。
+
+## 1. 数据集
+
+在 PaddleClas 的识别任务中，训练主体检测模型时主要用到了以下几个数据集。
+
+| 数据集       | 数据量   | 主体检测任务中使用的数据量   | 场景  | 数据集地址 |
+| :------------:  | :-------------: | :-------: | :-------: | :--------: |
+| Objects365 | 170W | 6k | 通用场景 | [地址](https://www.objects365.org/overview.html) |
+| COCO2017 | 12W | 5k  | 通用场景 | [地址](https://cocodataset.org/) |
+| iCartoonFace | 2k | 2k | 动漫人脸检测 | [地址](https://github.com/luxiangju-PersonAI/iCartoonFace) |
+| LogoDet-3k | 3k | 2k | Logo检测 | [地址](https://github.com/Wangjing1551/LogoDet-3K-Dataset) |
+| RPC | 3k | 3k  | 商品检测 | [地址](https://rpc-dataset.github.io/) |
+
+在实际训练的过程中，将所有数据集混合在一起。由于是主体检测，这里将所有标注出的检测框对应的类别都修改为 `前景` 的类别，最终融合的数据集中只包含 1 个类别，即前景。
+
+
+## 2. 模型选择
+
+目标检测方法种类繁多，比较常用的有两阶段检测器（如FasterRCNN系列等）；单阶段检测器（如YOLO、SSD等）；anchor-free检测器（如PicoDet、FCOS等）。PaddleDetection中针对服务端使用场景，自研了 PP-YOLO 系列模型；针对端侧（CPU和移动端等）使用场景，自研了 PicoDet 系列模型，在服务端和端侧均处于业界较为领先的水平。
+
+基于上述研究，PaddleClas 中提供了 2 个通用主体检测模型，为轻量级与服务端主体检测模型，分别适用于端侧场景以及服务端场景。下面的表格中给出了在上述 5 个数据集上的平均 mAP 以及它们的模型大小、预测速度对比信息。
+
+| 模型      | 模型结构   | 预训练模型下载地址   | inference模型下载地址  | mAP | inference模型大小(MB) | 单张图片预测耗时(不包含预处理)(ms) |
+| :------------:  | :-------------: | :------: | :-------: | :--------: | :-------: | :--------: |
+| 轻量级主体检测模型 | PicoDet | [地址](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/pretrain/picodet_PPLCNet_x2_5_mainbody_lite_v1.0_pretrained.pdparams) | [地址](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/inference/picodet_PPLCNet_x2_5_mainbody_lite_v1.0_infer.tar) | 40.1% | 30.1 | 29.8  |
+| 服务端主体检测模型 | PP-YOLOv2 | [地址](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/pretrain/ppyolov2_r50vd_dcn_mainbody_v1.0_pretrained.pdparams) | [地址](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/inference/ppyolov2_r50vd_dcn_mainbody_v1.0_infer.tar) | 42.5% | 210.5 | 466.6  |
+
+
+* 注意
+  * 速度评测机器的CPU具体信息为：`Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz`，速度指标为开启 mkldnn ，线程数设置为 10 测试得到。
+  * 主体检测的预处理过程较为耗时，平均每张图在上述机器上的时间在 40~55 ms 左右，没有包含在上述的预测耗时统计中。
+
+
+### 2.1 轻量级主体检测模型
+
+PicoDet 由 [PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection) 提出，是一个适用于CPU或者移动端场景的目标检测算法。具体地，它融合了下面一系列优化算法。
+
+- [ATSS](https://arxiv.org/abs/1912.02424)
+- [Generalized Focal Loss](https://arxiv.org/abs/2006.04388)
+- 余弦学习率策略
+- Cycle-EMA
+- 轻量级检测 head
+
+
+更多关于 PicoDet 的优化细节与 benchmark 可以参考 [PicoDet 系列模型介绍](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/picodet/README.md)。
+
+在轻量级主体检测任务中，为了更好地兼顾检测速度与效果，我们使用 PPLCNet_x2_5 作为主体检测模型的骨干网络，同时将训练与预测的图像尺度修改为了 640x640，其余配置与 [picodet_m_shufflenetv2_416_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/configs/picodet/picodet_m_shufflenetv2_416_coco.yml)完全一致。将数据集更换为自定义的主体检测数据集，进行训练，最终得到检测模型。
+
+
+### 2.2 服务端主体检测模型
+
+PP-YOLO 由 [PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection) 提出，从骨干网络、数据增广、正则化策略、损失函数、后处理等多个角度对 yolov3 模型进行深度优化，最终在"速度-精度"方面达到了业界领先的水平。具体地，优化的策略如下。
+
+- 更优的骨干网络: ResNet50vd-DCN
+- 更大的训练batch size: 8 GPUs，每GPU batch_size=24，对应调整学习率和迭代轮数
+- [Drop Block](https://arxiv.org/abs/1810.12890)
+- [Exponential Moving Average](https://www.investopedia.com/terms/e/ema.asp)
+- [IoU Loss](https://arxiv.org/pdf/1902.09630.pdf)
+- [Grid Sensitive](https://arxiv.org/abs/2004.10934)
+- [Matrix NMS](https://arxiv.org/pdf/2003.10152.pdf)
+- [CoordConv](https://arxiv.org/abs/1807.03247)
+- [Spatial Pyramid Pooling](https://arxiv.org/abs/1406.4729)
+- 更优的预训练模型
+
+更多关于 PP-YOLO 的详细介绍可以参考：[PP-YOLO 模型](https://github.com/PaddlePaddle/PaddleDetection/blob/release%2F2.1/configs/ppyolo/README_cn.md)。
+
+在服务端主体检测任务中，为了保证检测效果，我们使用 ResNet50vd-DCN 作为检测模型的骨干网络，使用配置文件 [ppyolov2_r50vd_dcn_365e_coco.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.1/configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml) ，更换为自定义的主体检测数据集，进行训练，最终得到检测模型。
+
+
+## 3. 模型训练
+
+本节主要介绍怎样基于PaddleDetection，基于自己的数据集，训练主体检测模型。
+
+### 3.1 环境准备
+
+下载PaddleDetection代码，安装requirements。
+
+```shell
+cd <path/to/clone/PaddleDetection>
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+
+cd PaddleDetection
+# 安装其他依赖
+pip install -r requirements.txt
+```
+
+更多安装教程，请参考: [安装文档](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.1/docs/tutorials/INSTALL_cn.md)
+
+### 3.2 数据准备
+
+对于自定义数据集，首先需要将自己的数据集修改为COCO格式，可以参考[自定义检测数据集教程](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.1/static/docs/tutorials/Custom_DataSet.md)制作COCO格式的数据集。
+
+主体检测任务中，所有的检测框均属于前景，在这里需要将标注文件中，检测框的`category_id`修改为1，同时将整个标注文件中的`categories`映射表修改为下面的格式，即整个类别映射表中只包含`前景`类别。
+
+```json
+[{u'id': 1, u'name': u'foreground', u'supercategory': u'foreground'}]
+```
+
+### 3.3 配置文件改动和说明
+
+我们使用 `configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml` 配置进行训练，配置文件摘要如下：
+
+<div align='center'>
+  <img src='../../images/det/PaddleDetection_config.png' width='400'/>
+</div>
+
+从上图看到 `ppyolov2_r50vd_dcn_365e_coco.yml` 配置需要依赖其他的配置文件，这些配置文件的含义如下:
+
+```
+coco_detection.yml：主要说明了训练数据和验证数据的路径
+
+runtime.yml：主要说明了公共的运行参数，比如是否使用GPU、每多少个epoch存储checkpoint等
+
+optimizer_365e.yml：主要说明了学习率和优化器的配置
+
+ppyolov2_r50vd_dcn.yml：主要说明模型和主干网络的情况
+
+ppyolov2_reader.yml：主要说明数据读取器配置，如 batch size，并发加载子进程数等，同时包含读取后预处理操作，如resize、数据增强等等
+```
+
+在主体检测任务中，需要将 `datasets/coco_detection.yml` 中的 `num_classes` 参数修改为 1 （只有 1 个前景类别），同时将训练集和测试集的路径修改为自定义数据集的路径。
+
+此外，也可以根据实际情况，修改上述文件，比如，如果显存溢出，可以将 batch size 和学习率等比缩小等。
+
+
+### 3.4 启动训练
+
+PaddleDetection 提供了单卡/多卡训练模式，满足用户多种训练需求。
+
+* GPU 单卡训练
+
+```bash
+# windows和Mac下不需要执行该命令
+export CUDA_VISIBLE_DEVICES=0
+python tools/train.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml
+```
+
+* GPU多卡训练
+
+```bash
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python -m paddle.distributed.launch --gpus 0,1,2,3 tools/train.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml --eval
+```
+
+--eval：表示边训练边验证。
+
+
+* (**推荐**)模型微调
+如果希望加载 PaddleClas 中已经训练好的主体检测模型，在自己的数据集上进行模型微调，可以使用下面的命令进行训练。
+
+```bash
+export CUDA_VISIBLE_DEVICES=0
+# 指定pretrain_weights参数，加载通用的主体检测预训练模型
+python tools/train.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml -o pretrain_weights=https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/rec/models/pretrain/ppyolov2_r50vd_dcn_mainbody_v1.0_pretrained.pdparams
+```
+
+* 模型恢复训练
+
+在日常训练过程中，有的用户由于一些原因导致训练中断，可以使用 `-r` 的命令恢复训练:
+
+```bash
+export CUDA_VISIBLE_DEVICES=0,1,2,3
+python -m paddle.distributed.launch --gpus 0,1,2,3 tools/train.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml --eval -r output/ppyolov2_r50vd_dcn_365e_coco/10000
+```
+
+注意：如果遇到 "`Out of memory error`" 问题, 尝试在 `ppyolov2_reader.yml` 文件中调小`batch_size`，同时等比例调小学习率。
+
+
+### 3.5 模型预测与调试
+
+使用下面的命令完成 PaddleDetection 的预测过程。
+
+```bash
+export CUDA_VISIBLE_DEVICES=0
+python tools/infer.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml --infer_img=your_image_path.jpg --output_dir=infer_output/ --draw_threshold=0.5 -o weights=output/ppyolov2_r50vd_dcn_365e_coco/model_final
+```
+
+`--draw_threshold` 是个可选参数. 根据 [NMS](https://ieeexplore.ieee.org/document/1699659) 的计算，不同阈值会产生不同的结果 `keep_top_k` 表示设置输出目标的最大数量，默认值为 100 ，用户可以根据自己的实际情况进行设定。
+
+### 3.6 模型导出与预测部署。
+
+执行导出模型脚本：
+
+```bash
+python tools/export_model.py -c configs/ppyolo/ppyolov2_r50vd_dcn_365e_coco.yml --output_dir=./inference -o weights=output/ppyolov2_r50vd_dcn_365e_coco/model_final.pdparams
+```
+
+预测模型会导出到 `inference/ppyolov2_r50vd_dcn_365e_coco` 目录下，分别为 `infer_cfg.yml` (预测不需要), `model.pdiparams`, `model.pdiparams.info`, `model.pdmodel` 。
+
+注意： `PaddleDetection` 导出的inference模型的文件格式为 `model.xxx`，这里如果希望与PaddleClas的inference模型文件格式保持一致，需要将其 `model.xxx` 文件修改为 `inference.xxx` 文件，用于后续主体检测的预测部署。
+
+更多模型导出教程，请参考： [EXPORT_MODEL](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.1/deploy/EXPORT_MODEL.md)
+
+最终，目录 `inference/ppyolov2_r50vd_dcn_365e_coco` 中包含 `inference.pdiparams`, `inference.pdiparams.info` 以及 `inference.pdmodel` 文件，其中 `inference.pdiparams` 为保存的 inference 模型权重文件， `inference.pdmodel` 为保存的 inference 模型结构文件。
+
+
+导出模型之后，在主体检测与识别任务中，就可以将检测模型的路径更改为该 inference 模型路径，完成预测。
+
+以商品识别为例，其配置文件为 [inference_product.yaml](../../../deploy/configs/inference_product.yaml) ，修改其中的 `Global.det_inference_model_dir` 字段为导出的主体检测 inference 模型目录，参考[图像识别快速开始教程](../tutorials/quick_start_recognition.md) ，即可完成商品检测与识别过程。
+
+
+### FAQ
+
+#### Q：可以使用其他的主体检测模型结构吗？
+
+* A：可以的，但是目前的检测预处理过程仅适配了 PicoDet 以及 YOLO 系列的预处理，因此在使用的时候，建议优先使用这两个系列的模型进行训练，如果希望使用 Faster RCNN 等其他系列的模型，需要按照 PaddleDetection 的数据预处理，修改下预处理逻辑，这块如果您有需求或者有问题的话，欢迎提 issue 或者在微信群里反馈。
+
+#### Q：可以修改主体检测的预测尺度吗？
+
+* A：可以的，但是需要注意 2 个地方
+  * PaddleClas 中提供的主体检测模型是基于 `640x640` 的分辨率去训练的，因此预测的时候也是默认使用 `640x640` 的分辨率进行预测，使用其他分辨率预测的话，精度会有所降低。
+  * 在模型导出的时候，建议也修改下模型导出的分辨率，保持模型导出、模型预测的分辨率一致。

ppcls/arch/backbone/__init__.py

@@ -58,6 +58,7 @@ from ppcls.arch.backbone.model_zoo.dla import DLA34, DLA46_c, DLA46x_c, DLA60, D
 from ppcls.arch.backbone.model_zoo.rednet import RedNet26, RedNet38, RedNet50, RedNet101, RedNet152
 from ppcls.arch.backbone.model_zoo.tnt import TNT_small
 from ppcls.arch.backbone.model_zoo.hardnet import HarDNet68, HarDNet85, HarDNet39_ds, HarDNet68_ds
+from ppcls.arch.backbone.model_zoo.cspnet import CSPDarkNet53
 from ppcls.arch.backbone.variant_models.resnet_variant import ResNet50_last_stage_stride1
 from ppcls.arch.backbone.variant_models.vgg_variant import VGG19Sigmoid
 

ppcls/arch/backbone/model_zoo/cspnet.py

+# copyright (c) 2021 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 import ParamAttr
+
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+    "CSPDarkNet53":
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/CSPDarkNet53_pretrained.pdparams"
+}
+
+MODEL_CFGS = {
+    "CSPDarkNet53": dict(
+        stem=dict(
+            out_chs=32, kernel_size=3, stride=1, pool=''),
+        stage=dict(
+            out_chs=(64, 128, 256, 512, 1024),
+            depth=(1, 2, 8, 8, 4),
+            stride=(2, ) * 5,
+            exp_ratio=(2., ) + (1., ) * 4,
+            bottle_ratio=(0.5, ) + (1.0, ) * 4,
+            block_ratio=(1., ) + (0.5, ) * 4,
+            down_growth=True, ))
+}
+
+__all__ = ['CSPDarkNet53'
+           ]  # model_registry will add each entrypoint fn to this
+
+
+class ConvBnAct(nn.Layer):
+    def __init__(self,
+                 input_channels,
+                 output_channels,
+                 kernel_size=1,
+                 stride=1,
+                 padding=None,
+                 dilation=1,
+                 groups=1,
+                 act_layer=nn.LeakyReLU,
+                 norm_layer=nn.BatchNorm2D):
+        super().__init__()
+        if padding is None:
+            padding = (kernel_size - 1) // 2
+        self.conv = nn.Conv2D(
+            in_channels=input_channels,
+            out_channels=output_channels,
+            kernel_size=kernel_size,
+            stride=stride,
+            padding=padding,
+            dilation=dilation,
+            groups=groups,
+            weight_attr=ParamAttr(),
+            bias_attr=False)
+
+        self.bn = norm_layer(num_features=output_channels)
+        self.act = act_layer()
+
+    def forward(self, inputs):
+        x = self.conv(inputs)
+        x = self.bn(x)
+        if self.act is not None:
+            x = self.act(x)
+        return x
+
+
+def create_stem(in_chans=3,
+                out_chs=32,
+                kernel_size=3,
+                stride=2,
+                pool='',
+                act_layer=None,
+                norm_layer=None):
+    stem = nn.Sequential()
+    if not isinstance(out_chs, (tuple, list)):
+        out_chs = [out_chs]
+    assert len(out_chs)
+    in_c = in_chans
+    for i, out_c in enumerate(out_chs):
+        conv_name = f'conv{i + 1}'
+        stem.add_sublayer(
+            conv_name,
+            ConvBnAct(
+                in_c,
+                out_c,
+                kernel_size,
+                stride=stride if i == 0 else 1,
+                act_layer=act_layer,
+                norm_layer=norm_layer))
+        in_c = out_c
+        last_conv = conv_name
+    if pool:
+        stem.add_sublayer(
+            'pool', nn.MaxPool2D(
+                kernel_size=3, stride=2, padding=1))
+    return stem, dict(
+        num_chs=in_c, reduction=stride, module='.'.join(['stem', last_conv]))
+
+
+class DarkBlock(nn.Layer):
+    def __init__(self,
+                 in_chs,
+                 out_chs,
+                 dilation=1,
+                 bottle_ratio=0.5,
+                 groups=1,
+                 act_layer=nn.ReLU,
+                 norm_layer=nn.BatchNorm2D,
+                 attn_layer=None,
+                 drop_block=None):
+        super(DarkBlock, self).__init__()
+        mid_chs = int(round(out_chs * bottle_ratio))
+        ckwargs = dict(act_layer=act_layer, norm_layer=norm_layer)
+        self.conv1 = ConvBnAct(in_chs, mid_chs, kernel_size=1, **ckwargs)
+        self.conv2 = ConvBnAct(
+            mid_chs,
+            out_chs,
+            kernel_size=3,
+            dilation=dilation,
+            groups=groups,
+            **ckwargs)
+
+    def forward(self, x):
+        shortcut = x
+        x = self.conv1(x)
+        x = self.conv2(x)
+        x = x + shortcut
+        return x
+
+
+class CrossStage(nn.Layer):
+    def __init__(self,
+                 in_chs,
+                 out_chs,
+                 stride,
+                 dilation,
+                 depth,
+                 block_ratio=1.,
+                 bottle_ratio=1.,
+                 exp_ratio=1.,
+                 groups=1,
+                 first_dilation=None,
+                 down_growth=False,
+                 cross_linear=False,
+                 block_dpr=None,
+                 block_fn=DarkBlock,
+                 **block_kwargs):
+        super(CrossStage, self).__init__()
+        first_dilation = first_dilation or dilation
+        down_chs = out_chs if down_growth else in_chs
+        exp_chs = int(round(out_chs * exp_ratio))
+        block_out_chs = int(round(out_chs * block_ratio))
+        conv_kwargs = dict(
+            act_layer=block_kwargs.get('act_layer'),
+            norm_layer=block_kwargs.get('norm_layer'))
+
+        if stride != 1 or first_dilation != dilation:
+            self.conv_down = ConvBnAct(
+                in_chs,
+                down_chs,
+                kernel_size=3,
+                stride=stride,
+                dilation=first_dilation,
+                groups=groups,
+                **conv_kwargs)
+            prev_chs = down_chs
+        else:
+            self.conv_down = None
+            prev_chs = in_chs
+
+        self.conv_exp = ConvBnAct(
+            prev_chs, exp_chs, kernel_size=1, **conv_kwargs)
+        prev_chs = exp_chs // 2  # output of conv_exp is always split in two
+
+        self.blocks = nn.Sequential()
+        for i in range(depth):
+            self.blocks.add_sublayer(
+                str(i),
+                block_fn(prev_chs, block_out_chs, dilation, bottle_ratio,
+                         groups, **block_kwargs))
+            prev_chs = block_out_chs
+
+        # transition convs
+        self.conv_transition_b = ConvBnAct(
+            prev_chs, exp_chs // 2, kernel_size=1, **conv_kwargs)
+        self.conv_transition = ConvBnAct(
+            exp_chs, out_chs, kernel_size=1, **conv_kwargs)
+
+    def forward(self, x):
+        if self.conv_down is not None:
+            x = self.conv_down(x)
+        x = self.conv_exp(x)
+        split = x.shape[1] // 2
+        xs, xb = x[:, :split], x[:, split:]
+        xb = self.blocks(xb)
+        xb = self.conv_transition_b(xb)
+        out = self.conv_transition(paddle.concat([xs, xb], axis=1))
+        return out
+
+
+class DarkStage(nn.Layer):
+    def __init__(self,
+                 in_chs,
+                 out_chs,
+                 stride,
+                 dilation,
+                 depth,
+                 block_ratio=1.,
+                 bottle_ratio=1.,
+                 groups=1,
+                 first_dilation=None,
+                 block_fn=DarkBlock,
+                 block_dpr=None,
+                 **block_kwargs):
+        super().__init__()
+        first_dilation = first_dilation or dilation
+
+        self.conv_down = ConvBnAct(
+            in_chs,
+            out_chs,
+            kernel_size=3,
+            stride=stride,
+            dilation=first_dilation,
+            groups=groups,
+            act_layer=block_kwargs.get('act_layer'),
+            norm_layer=block_kwargs.get('norm_layer'))
+
+        prev_chs = out_chs
+        block_out_chs = int(round(out_chs * block_ratio))
+        self.blocks = nn.Sequential()
+        for i in range(depth):
+            self.blocks.add_sublayer(
+                str(i),
+                block_fn(prev_chs, block_out_chs, dilation, bottle_ratio,
+                         groups, **block_kwargs))
+            prev_chs = block_out_chs
+
+    def forward(self, x):
+        x = self.conv_down(x)
+        x = self.blocks(x)
+        return x
+
+
+def _cfg_to_stage_args(cfg, curr_stride=2, output_stride=32):
+    # get per stage args for stage and containing blocks, calculate strides to meet target output_stride
+    num_stages = len(cfg['depth'])
+    if 'groups' not in cfg:
+        cfg['groups'] = (1, ) * num_stages
+    if 'down_growth' in cfg and not isinstance(cfg['down_growth'],
+                                               (list, tuple)):
+        cfg['down_growth'] = (cfg['down_growth'], ) * num_stages
+    stage_strides = []
+    stage_dilations = []
+    stage_first_dilations = []
+    dilation = 1
+    for cfg_stride in cfg['stride']:
+        stage_first_dilations.append(dilation)
+        if curr_stride >= output_stride:
+            dilation *= cfg_stride
+            stride = 1
+        else:
+            stride = cfg_stride
+            curr_stride *= stride
+        stage_strides.append(stride)
+        stage_dilations.append(dilation)
+    cfg['stride'] = stage_strides
+    cfg['dilation'] = stage_dilations
+    cfg['first_dilation'] = stage_first_dilations
+    stage_args = [
+        dict(zip(cfg.keys(), values)) for values in zip(*cfg.values())
+    ]
+    return stage_args
+
+
+class CSPNet(nn.Layer):
+    def __init__(self,
+                 cfg,
+                 in_chans=3,
+                 class_num=1000,
+                 output_stride=32,
+                 global_pool='avg',
+                 drop_rate=0.,
+                 act_layer=nn.LeakyReLU,
+                 norm_layer=nn.BatchNorm2D,
+                 zero_init_last_bn=True,
+                 stage_fn=CrossStage,
+                 block_fn=DarkBlock):
+        super().__init__()
+        self.class_num = class_num
+        self.drop_rate = drop_rate
+        assert output_stride in (8, 16, 32)
+        layer_args = dict(act_layer=act_layer, norm_layer=norm_layer)
+
+        # Construct the stem
+        self.stem, stem_feat_info = create_stem(in_chans, **cfg['stem'],
+                                                **layer_args)
+        self.feature_info = [stem_feat_info]
+        prev_chs = stem_feat_info['num_chs']
+        curr_stride = stem_feat_info[
+            'reduction']  # reduction does not include pool
+        if cfg['stem']['pool']:
+            curr_stride *= 2
+
+        # Construct the stages
+        per_stage_args = _cfg_to_stage_args(
+            cfg['stage'], curr_stride=curr_stride, output_stride=output_stride)
+        self.stages = nn.LayerList()
+        for i, sa in enumerate(per_stage_args):
+            self.stages.add_sublayer(
+                str(i),
+                stage_fn(
+                    prev_chs, **sa, **layer_args, block_fn=block_fn))
+            prev_chs = sa['out_chs']
+            curr_stride *= sa['stride']
+            self.feature_info += [
+                dict(
+                    num_chs=prev_chs,
+                    reduction=curr_stride,
+                    module=f'stages.{i}')
+            ]
+
+        # Construct the head
+        self.num_features = prev_chs
+
+        self.pool = nn.AdaptiveAvgPool2D(1)
+        self.flatten = nn.Flatten(1)
+        self.fc = nn.Linear(
+            prev_chs,
+            class_num,
+            weight_attr=ParamAttr(),
+            bias_attr=ParamAttr())
+
+    def forward(self, x):
+        x = self.stem(x)
+        for stage in self.stages:
+            x = stage(x)
+        x = self.pool(x)
+        x = self.flatten(x)
+        x = self.fc(x)
+        return x
+
+
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    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 CSPDarkNet53(pretrained=False, use_ssld=False, **kwargs):
+    model = CSPNet(MODEL_CFGS["CSPDarkNet53"], block_fn=DarkBlock, **kwargs)
+    _load_pretrained(
+        pretrained, model, MODEL_URLS["CSPDarkNet53"], use_ssld=use_ssld)
+    return model

ppcls/configs/ImageNet/CSPNet/CSPDarkNet53.yaml

+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  save_interval: 1
+  eval_during_train: True
+  eval_interval: 1
+  epochs: 120
+  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
+
+# model architecture
+Arch:
+  name: CSPDarkNet53
+  class_num: 1000
+ 
+# loss function config for traing/eval process
+Loss:
+  Train:
+    - CELoss:
+        weight: 1.0
+  Eval:
+    - CELoss:
+        weight: 1.0
+
+
+Optimizer:
+  name: Momentum
+  momentum: 0.9
+  lr:
+    name: Piecewise
+    decay_epochs: [30, 60, 90]
+    values: [0.1, 0.01, 0.001, 0.0001]
+  regularizer:
+    name: 'L2'
+    coeff: 0.0001
+
+
+# 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: 256
+        - RandFlipImage:
+            flip_code: 1
+        - 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: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      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: 288
+        - CropImage:
+            size: 256
+        - 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: 64
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/whl/demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 288
+    - CropImage:
+        size: 256
+    - 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]

ppcls/data/preprocess/ops/operators.py

@@ -60,7 +60,7 @@ class UnifiedResize(object):
             if isinstance(interpolation, str):
                 interpolation = _cv2_interp_from_str[interpolation.lower()]
             # compatible with opencv < version 4.4.0
-            elif not interpolation:
+            elif interpolation is None:
                 interpolation = cv2.INTER_LINEAR
             self.resize_func = partial(cv2.resize, interpolation=interpolation)
         elif backend.lower() == "pil":

ppcls/optimizer/__init__.py

@@ -41,7 +41,8 @@ def build_lr_scheduler(lr_config, epochs, step_each_epoch):
     return lr
 
 
-def build_optimizer(config, epochs, step_each_epoch, model_list):
+# model_list is None in static graph
+def build_optimizer(config, epochs, step_each_epoch, model_list=None):
     config = copy.deepcopy(config)
     # step1 build lr
     lr = build_lr_scheduler(config.pop('lr'), epochs, step_each_epoch)

ppcls/optimizer/optimizer.py

@@ -18,6 +18,8 @@ from __future__ import print_function
 
 from paddle import optimizer as optim
 
+from ppcls.utils import logger
+
 
 class Momentum(object):
     """
@@ -43,7 +45,9 @@ class Momentum(object):
         self.multi_precision = multi_precision
 
     def __call__(self, model_list):
-        parameters = sum([m.parameters() for m in model_list], [])
+        # model_list is None in static graph
+        parameters = sum([m.parameters() for m in model_list],
+                         []) if model_list else None
         opt = optim.Momentum(
             learning_rate=self.learning_rate,
             momentum=self.momentum,
@@ -79,7 +83,9 @@ class Adam(object):
         self.multi_precision = multi_precision
 
     def __call__(self, model_list):
-        parameters = sum([m.parameters() for m in model_list], [])
+        # model_list is None in static graph
+        parameters = sum([m.parameters() for m in model_list],
+                         []) if model_list else None
         opt = optim.Adam(
             learning_rate=self.learning_rate,
             beta1=self.beta1,
@@ -123,7 +129,9 @@ class RMSProp(object):
         self.grad_clip = grad_clip
 
     def __call__(self, model_list):
-        parameters = sum([m.parameters() for m in model_list], [])
+        # model_list is None in static graph
+        parameters = sum([m.parameters() for m in model_list],
+                         []) if model_list else None
         opt = optim.RMSProp(
             learning_rate=self.learning_rate,
             momentum=self.momentum,
@@ -160,18 +168,28 @@ class AdamW(object):
         self.one_dim_param_no_weight_decay = one_dim_param_no_weight_decay
 
     def __call__(self, model_list):
-        parameters = sum([m.parameters() for m in model_list], [])
+        # model_list is None in static graph
+        parameters = sum([m.parameters() for m in model_list],
+                         []) if model_list else None
+
+        # TODO(gaotingquan): model_list is None when in static graph, "no_weight_decay" not work.
+        if model_list is None:
+            if self.one_dim_param_no_weight_decay or len(
+                    self.no_weight_decay_name_list) != 0:
+                msg = "\"AdamW\" does not support setting \"no_weight_decay\" in static graph. Please use dynamic graph."
+                logger.error(Exception(msg))
+                raise Exception(msg)
 
         self.no_weight_decay_param_name_list = [
             p.name for model in model_list for n, p in model.named_parameters()
             if any(nd in n for nd in self.no_weight_decay_name_list)
-        ]
+        ] if model_list else []
 
         if self.one_dim_param_no_weight_decay:
             self.no_weight_decay_param_name_list += [
                 p.name for model in model_list
                 for n, p in model.named_parameters() if len(p.shape) == 1
-            ]
+            ] if model_list else []
 
         opt = optim.AdamW(
             learning_rate=self.learning_rate,

tests/config/DarkNet53.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/GhostNet_x0_5.txt

+===========================train_params===========================
+model_name:GhostNet_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/GhostNet_x1_0.txt

+===========================train_params===========================
+model_name:GhostNet_x1_0
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_0.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_0.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x1_0_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/GhostNet_x1_3.txt

+===========================train_params===========================
+model_name:GhostNet_x1_3
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_3.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_3.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/GhostNet/GhostNet_x1_3.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x1_3_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/HRNet_W18_C.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/LeViT_128S.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|Fasle
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/MobileNetV1.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/MobileNetV1_x0_25.txt

+===========================train_params===========================
+model_name:MobileNetV1_x0_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV1_x0_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV1_x0_5.txt

+===========================train_params===========================
+model_name:MobileNetV1_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV1_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV1_x0_75.txt

+===========================train_params===========================
+model_name:MobileNetV1_x0_75
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV1_x0_75_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV2.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/MobileNetV2_x0_25.txt

+===========================train_params===========================
+model_name:MobileNetV2_x0_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV2_x0_5.txt

+===========================train_params===========================
+model_name:MobileNetV2_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV2_x0_75.txt

+===========================train_params===========================
+model_name:MobileNetV2_x0_75
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_75.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_75.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x0_75.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_75_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV2_x1_5.txt

+===========================train_params===========================
+model_name:MobileNetV2_x1_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x1_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x1_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x1_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x1_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV2_x2_0.txt

+===========================train_params===========================
+model_name:MobileNetV2_x2_0
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x2_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x2_0.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV2/MobileNetV2_x2_0.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x2_0_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_large_x0_35.txt

+===========================train_params===========================
+model_name:MobileNetV3_large_x0_35
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_large_x0_35_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_large_x0_5.txt

+===========================train_params===========================
+model_name:MobileNetV3_large_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_large_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_large_x0_75.txt

+===========================train_params===========================
+model_name:MobileNetV3_large_x0_75
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_large_x0_75_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_large_x1_0.txt

@@ -31,17 +31,17 @@ norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNe
 quant_export:tools/export_model.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_quantization.yaml
 fpgm_export:tools/export_model.py -c ppcls/configs/slim/MobileNetV3_large_x1_0_prune.yaml
 distill_export:null
-export1:null
+kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Global.save_inference_dir=./inference
 export2:null
 inference_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/whole_chain/MobileNetV3_large_x1_0_inference.tar
 infer_model:../inference/
-kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml -o Global.save_inference_dir=./inference
+infer_export:null
 infer_quant:Fasle
 inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/MobileNetV3_large_x1_25.txt

+===========================train_params===========================
+model_name:MobileNetV3_large_x1_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_large_x1_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_small_x0_35.txt

+===========================train_params===========================
+model_name:MobileNetV3_small_x0_35
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_small_x0_35_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_small_x0_5.txt

+===========================train_params===========================
+model_name:MobileNetV3_small_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_small_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_small_x0_75.txt

+===========================train_params===========================
+model_name:MobileNetV3_small_x0_75
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_small_x0_75_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_small_x1_0.txt

+===========================train_params===========================
+model_name:MobileNetV3_small_x1_0
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_small_x1_0_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/MobileNetV3_small_x1_25.txt

+===========================train_params===========================
+model_name:MobileNetV3_small_x1_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/MobileNetV3_small_x1_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/PPLCNet_x0_25.txt

+===========================train_params===========================
+model_name:PPLCNet_x0_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPLCNet_x0_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/PPLCNet_x0_35.txt

+===========================train_params===========================
+model_name:PPLCNet_x0_35
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_35.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_35.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_35.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPLCNet_x0_35_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/PPLCNet_x0_5.txt

+===========================train_params===========================
+model_name:PPLCNet_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPLCNet_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/PPLCNet_x1_0.txt

+===========================train_params===========================
+model_name:PPLCNet_x1_0
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/legendary_models/PPLCNet_x1_0_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/ResNeXt101_vd_64x4d.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/ResNet50_vd.txt

@@ -31,17 +31,17 @@ norm_export:tools/export_model.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.y
 quant_export:tools/export_model.py -c ppcls/configs/slim/ResNet50_vd_quantization.yaml
 fpgm_export:tools/export_model.py -c ppcls/configs/slim/ResNet50_vd_prune.yaml
 distill_export:null
-export1:null
+kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml -o Global.save_inference_dir=./inference
 export2:null
 infer_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/whole_chain/ResNet50_vd_inference.tar
 infer_model:../inference/
-kl_quant:deploy/slim/quant_post_static.py -c ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml -o Global.save_inference_dir=./inference
+infer_export:null
 infer_quant:Fasle
 inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/ShuffleNetV2_x0_25.txt

+===========================train_params===========================
+model_name:ShuffleNetV2_x0_25
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_25.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_25.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_25.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_25_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/ShuffleNetV2_x0_33.txt

+===========================train_params===========================
+model_name:ShuffleNetV2_x0_33
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_33.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_33.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_33.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_33_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/ShuffleNetV2_x0_5.txt

+===========================train_params===========================
+model_name:ShuffleNetV2_x0_5
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_infer=2|whole_train_infer=120
+-o Global.output_dir:./output/
+-o DataLoader.Train.sampler.batch_size:8
+-o Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./dataset/ILSVRC2012/val
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_5.yaml -o Global.seed=1234 -o DataLoader.Train.sampler.shuffle=False -o DataLoader.Train.loader.num_workers=0 -o DataLoader.Train.loader.use_shared_memory=False
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_5.yaml
+null:null
+##
+===========================infer_params==========================
+-o Global.save_inference_dir:./inference
+-o Global.pretrained_model:
+norm_export:tools/export_model.py -c ppcls/configs/ImageNet/ShuffleNet/ShuffleNetV2_x0_5.yaml
+quant_export:null
+fpgm_export:null
+distill_export:null
+kl_quant:null
+export2:null
+pretrained_model_url:https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_5_pretrained.pdparams
+infer_model:../inference/
+infer_export:True
+infer_quant:Fasle
+inference:python/predict_cls.py -c configs/inference_cls.yaml
+-o Global.use_gpu:True|False
+-o Global.enable_mkldnn:True|False
+-o Global.cpu_num_threads:1|6
+-o Global.batch_size:1|16
+-o Global.use_tensorrt:True|False
+-o Global.use_fp16:True|False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val
+-o Global.save_log_path:null
+-o Global.benchmark:True
+null:null

tests/config/ShuffleNetV2_x1_0.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/config/SwinTransformer_tiny_patch4_window7_224.txt

@@ -41,7 +41,7 @@ inference:python/predict_cls.py -c configs/inference_cls.yaml
 -o Global.use_gpu:True|False
 -o Global.enable_mkldnn:True|False
 -o Global.cpu_num_threads:1|6
--o Global.batch_size:1
+-o Global.batch_size:1|16
 -o Global.use_tensorrt:True|False
 -o Global.use_fp16:True|False
 -o Global.inference_model_dir:../inference

tests/prepare.sh

@@ -7,6 +7,15 @@ dataline=$(cat ${FILENAME})
 # parser params
 IFS=$'\n'
 lines=(${dataline})
+
+function func_parser_key(){
+    strs=$1
+    IFS=":"
+    array=(${strs})
+    tmp=${array[0]}
+    echo ${tmp}
+}
+
 function func_parser_value(){
     strs=$1
     IFS=":"
@@ -20,7 +29,8 @@ function func_parser_value(){
     fi
 }
 model_name=$(func_parser_value "${lines[1]}")
-inference_model_url=$(func_parser_value "${lines[35]}")
+model_url_value=$(func_parser_value "${lines[35]}")
+model_url_key=$(func_parser_key "${lines[35]}")
 
 if [ ${MODE} = "lite_train_infer" ] || [ ${MODE} = "whole_infer" ];then
     # pretrain lite train data
@@ -44,9 +54,12 @@ elif [ ${MODE} = "infer" ] || [ ${MODE} = "cpp_infer" ];then
     mv val.txt val_list.txt
     ln -s val_list.txt train_list.txt
     cd ../../
-    # download inference model
-    eval "wget -nc $inference_model_url"
-    tar xf "${model_name}_inference.tar"
+    # download inference or pretrained model
+    eval "wget -nc $model_url_value"
+    if [[ $model_url_key == *inference* ]]; then
+	rm -rf inference
+	tar xf "${model_name}_inference.tar"
+    fi
 
 elif [ ${MODE} = "whole_train_infer" ];then
     cd dataset

tests/test.sh

@@ -113,14 +113,14 @@ norm_export=$(func_parser_value "${lines[29]}")
 pact_export=$(func_parser_value "${lines[30]}")
 fpgm_export=$(func_parser_value "${lines[31]}")
 distill_export=$(func_parser_value "${lines[32]}")
-export_key1=$(func_parser_key "${lines[33]}")
-export_value1=$(func_parser_value "${lines[33]}")
+kl_quant_cmd_key=$(func_parser_key "${lines[33]}")
+kl_quant_cmd_value=$(func_parser_value "${lines[33]}")
 export_key2=$(func_parser_key "${lines[34]}")
 export_value2=$(func_parser_value "${lines[34]}")
 
 # parser inference model 
 infer_model_dir_list=$(func_parser_value "${lines[36]}")
-infer_export_list=$(func_parser_value "${lines[37]}")
+infer_export_flag=$(func_parser_value "${lines[37]}")
 infer_is_quant=$(func_parser_value "${lines[38]}")
 # parser inference 
 inference_py=$(func_parser_value "${lines[39]}")
@@ -295,8 +295,13 @@ if [ ${MODE} = "infer" ]; then
     eval $env
     export Count=0
     IFS="|"
-    infer_run_exports=(${infer_export_list})
+    infer_export_flag=(${infer_export_flag})
     infer_quant_flag=(${infer_is_quant})
+    if [ ${infer_export_flag} != "null" ]  && [ ${infer_export_flag} != "False" ]; then
+	rm -rf ${infer_model_dir_list/..\//}
+        export_cmd="${python} ${norm_export} -o Global.pretrained_model=${model_name}_pretrained -o Global.save_inference_dir=${infer_model_dir_list/..\//}"
+        eval $export_cmd
+    fi
     cd deploy
     for infer_model in ${infer_model_dir_list[*]}; do
         #run inference
@@ -308,9 +313,9 @@ if [ ${MODE} = "infer" ]; then
     cd ..
 
     # for kl_quant
-    echo "kl_quant"
-    if [ ${infer_run_exports} ]; then
-	command="${python} ${infer_run_exports}"
+    if [ ${kl_quant_cmd_value} != "null" ] && [ ${kl_quant_cmd_value} != "False" ]; then
+	echo "kl_quant"
+	command="${python} ${kl_quant_cmd_value}"
 	eval $command
 	last_status=${PIPESTATUS[0]}
 	status_check $last_status "${command}" "${status_log}"