add docs

docs/zh_CN/models/ImageNet1k/README.md

@@ -43,7 +43,7 @@
   - [4.1 服务器端模型](#Transformer_server)
     - [ViT 系列](#ViT)
     - [DeiT 系列](#DeiT)
-    - [SwinTransformer 系列](#SwinTransformer)
+    - [SwinTransformer & SwinTransformerV2系列](#SwinTransformer)
     - [Twins 系列](#Twins)
     - [CSwinTransformer 系列](#CSwinTransformer)
     - [PVTV2 系列](#PVTV2)
@@ -636,7 +636,7 @@ DeiT（Data-efficient Image Transformers）系列模型的精度、速度指标
 
 <a name="SwinTransformer"></a>
 
-## SwinTransformer 系列 <sup>[[27](#ref27)]</sup>
+## SwinTransformer  & SwinTransformerV2 系列 <sup>[[27](#ref27)]</sup><sup>[[50](#ref50)]</sup>
 
 关于 SwinTransformer 系列模型的精度、速度指标如下表所示，更多介绍可以参考：[SwinTransformer 系列模型文档](SwinTransformer.md)。
 
@@ -651,6 +651,19 @@ DeiT（Data-efficient Image Transformers）系列模型的精度、速度指标
 | SwinTransformer_large_patch4_window7_224<sup>[1]</sup>    | 0.8596 | 0.9783 | 15.74 | 38.57 | 71.49 | 34.02 | 196.43  | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_large_patch4_window7_224_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformer_large_patch4_window7_224_22kto1k_infer.tar) |
 | SwinTransformer_large_patch4_window12_384<sup>[1]</sup>   | 0.8719 | 0.9823 | 32.61 | 116.59 | 223.23 | 99.97 | 196.43 | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_large_patch4_window12_384_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformer_large_patch4_window12_384_22kto1k_infer.tar) |
 
+| 模型       | Top-1 Acc | Top-5 Acc | time(ms)<br>bs=1 | time(ms)<br>bs=4 | time(ms)<br/>bs=8 | FLOPs(G) | Params(M) | 预训练模型下载地址                                               | inference模型下载地址                                      |
+| ---------- | --------- | --------- | ---------------- | ---------------- | -------- | --------- | ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
+| SwinTransformerV2_tiny_patch4_window8_256    | 0.8177 | 0.9588 | - | - | - | 4.34  | 21.87   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_tiny_patch4_window8_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_tiny_patch4_window8_256_infer.tar) |
+| SwinTransformerV2_tiny_patch4_window16_256   | 0.8283 | 0.9623 | - | - | - | 4.38  | 21.87   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_tiny_patch4_window16_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_tiny_patch4_window16_256_infer.tar) |
+| SwinTransformerV2_small_patch4_window8_256    | 0.8373 | 0.9662 | - | - | - | 8.44 | 37.93   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_small_patch4_window8_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_small_patch4_window8_256_infer.tar) |
+| SwinTransformerV2_small_patch4_window16_256   | 0.8414 | 0.9681 | - | - | - | 8.54 | 37.93   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_small_patch4_window16_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_small_patch4_window16_256_infer.tar) |
+| SwinTransformerV2_base_patch4_window8_256    | 0.8419 | 0.9687 | - | - | - | 14.97 | 66.96   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_base_patch4_window8_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_base_patch4_window8_256_infer.tar) |
+| SwinTransformerV2_base_patch4_window16_256   | 0.8458 | 0.9706 | - | - | - | 15.11 | 66.96   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_base_patch4_window16_256_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_base_patch4_window16_256_infer.tar) |
+| SwinTransformerV2_base_patch4_window12to16_256<sup>[1]</sup>     | 0.8616 | 0.9789 | - | - | - | 15.11 | 66.96   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_base_patch4_window12to16_256_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_base_patch4_window12to16_256_22kto1k_infer.tar) |
+| SwinTransformerV2_base_patch4_window24_384<sup>[1]</sup>    | 0.8714 | 0.9824 | - | - | - | 34.00 | 66.96   | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_base_patch4_window24_384_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_base_patch4_window24_384_22kto1k_infer.tar) |
+| SwinTransformerV2_large_patch4_window16_256<sup>[1]</sup>    | 0.8689 | 0.9804 | - | - | - | 33.82 | 149.59  | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_large_patch4_window16_256_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_large_patch4_window16_256_22kto1k_infer.tar) |
+| SwinTransformerV2_large_patch4_window24_384<sup>[1]</sup>   | 0.8747 | 0.9827 | - | - | - | 76.12 | 149.59 | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformerV2_large_patch4_window24_384_22kto1k_pretrained.pdparams) | [下载链接](https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/inference/SwinTransformerV2_large_patch4_window24_384_22kto1k_infer.tar) |
+
 [1]：基于 ImageNet22k 数据集预训练，然后在 ImageNet1k 数据集迁移学习得到。
 
 <a name="Twins"></a>
@@ -894,3 +907,5 @@ TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE.
 <a name="ref48">[48]</a>Kunchang Li, Yali Wang, Junhao Zhang, Peng Gao, Guanglu Song, Yu Liu, Hongsheng Li, Yu Qiao. UniFormer: Unifying Convolution and Self-attention for Visual Recognition
 
 <a name="ref49">[49]</a>Mingyuan Mao, Renrui Zhang, Honghui Zheng, Peng Gao, Teli Ma, Yan Peng, Errui Ding, Baochang Zhang, Shumin Han. Dual-stream Network for Visual Recognition.
+
+<a name="ref50">[50]</a>Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo. Swin Transformer V2: Scaling Up Capacity and Resolution

docs/zh_CN/models/ImageNet1k/SwinTransformerV2.md

+# SwinTransformerV2
+-----
+
+## 目录
+
+- [1. 模型介绍](#1)
+    - [1.1 模型简介](#1.1)
+    - [1.2 模型指标](#1.2)
+    - [1.3 Benchmark](#1.3)
+      - [1.3.1 基于 V100 GPU 的预测速度](#1.3.1)
+- [2. 模型快速体验](#2)
+- [3. 模型训练、评估和预测](#3)
+- [4. 模型推理部署](#4)
+  - [4.1 推理模型准备](#4.1)
+  - [4.2 基于 Python 预测引擎推理](#4.2)
+  - [4.3 基于 C++ 预测引擎推理](#4.3)
+  - [4.4 服务化部署](#4.4)
+  - [4.5 端侧部署](#4.5)
+  - [4.6 Paddle2ONNX 模型转换与预测](#4.6)
+
+
+<a name='1'></a>
+
+## 1. 模型介绍
+
+<a name='1.1'></a>
+
+### 1.1 模型简介
+
+SwinTransformerV2 在 SwinTransformer 的基础上进行改进，可处理大尺寸图像。通过提升模型容量与输入分辨率，SwinTransformerV2 在四个代表性基准数据集上取得了新记录。[论文地址](https://arxiv.org/abs/2111.09883)。
+
+<a name='1.2'></a>
+
+### 1.2 模型指标
+
+| Models           | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPs<br>(G) | Params<br>(M) |
+|:--:|:--:|:--:|:--:|:--:|:--:|:--:|
+| SwinTransformerV2_tiny_patch4_window8_256    | 0.8177 | 0.9588 | 0.818 | 0.959 | 4.3  | 21.9   |
+| SwinTransformerV2_tiny_patch4_window16_256   | 0.8283 | 0.9623 | 0.828 | 0.962 | 4.4  | 21.9   |
+| SwinTransformerV2_small_patch4_window8_256    | 0.8373 | 0.9662 | 0.837 | 0.966 | 8.4 | 37.9   |
+| SwinTransformerV2_small_patch4_window16_256   | 0.8414 | 0.9681 | 0.841 | 0.968 | 8.5 | 37.9   |
+| SwinTransformerV2_base_patch4_window8_256    | 0.8419 | 0.9687 | 0.842 | 0.969 | 15.0 | 67.0   |
+| SwinTransformerV2_base_patch4_window16_256   | 0.8458 | 0.9706 | 0.846 | 0.970 | 15.1 | 67.0   |
+| SwinTransformerV2_base_patch4_window12to16_256<sup>[1]</sup>    | 0.8616 | 0.9789 | 0.862 | 0.979 | 15.1 | 67.0   |
+| SwinTransformerV2_base_patch4_window24_384<sup>[1]</sup>   | 0.8714 | 0.9824 | 0.871 | 0.982 | 34.0 | 67.0   |
+| SwinTransformerV2_large_patch4_window16_256<sup>[1]</sup>   | 0.8689 | 0.9804 | 0.869 | 0.980 | 33.8 | 149.6 |
+| SwinTransformerV2_large_patch4_window24_384<sup>[1]</sup>  | 0.8747 | 0.9827 | 0.876 | 0.983 | 76.1 | 149.6 |
+
+[1]：基于 ImageNet22k 数据集预训练，然后在 ImageNet1k 数据集迁移学习得到。
+
+**备注：**
+1. 与 Reference 的精度差异源于数据预处理不同。
+2. PaddleClas 所提供的该系列模型的预训练模型权重，均是基于其官方提供的权重转得。
+
+<a name='1.3'></a>
+
+### 1.3 Benchmark
+
+<a name='1.3.1'></a>
+
+#### 1.3.1 基于 V100 GPU 的预测速度
+
+敬请期待
+
+<a name="2"></a>  
+
+## 2. 模型快速体验
+
+安装 paddlepaddle 和 paddleclas 即可快速对图片进行预测，体验方法可以参考[ResNet50 模型快速体验](./ResNet.md#2)。
+
+<a name="3"></a>
+
+## 3. 模型训练、评估和预测
+
+此部分内容包括训练环境配置、ImageNet数据的准备、SwinTransformer 在 ImageNet 上的训练、评估、预测等内容。在 `ppcls/configs/ImageNet/SwinTransformer/` 中提供了 SwinTransformer 的训练配置，可以通过如下脚本启动训练：此部分内容可以参考[ResNet50 模型训练、评估和预测](./ResNet.md#3)。
+
+**备注：** 由于 SwinTransformer 系列模型默认使用的 GPU 数量为 8 个，所以在训练时，需要指定8个GPU，如`python3 -m paddle.distributed.launch --gpus="0,1,2,3,4,5,6,7" tools/train.py -c xxx.yaml`, 如果使用 4 个 GPU 训练，默认学习率需要减小一半，精度可能有损。
+
+<a name="4"></a>
+
+## 4. 模型推理部署
+
+<a name="4.1"></a>
+
+### 4.1 推理模型准备
+
+Paddle Inference 是飞桨的原生推理库， 作用于服务器端和云端，提供高性能的推理能力。相比于直接基于预训练模型进行预测，Paddle Inference可使用 MKLDNN、CUDNN、TensorRT 进行预测加速，从而实现更优的推理性能。更多关于Paddle Inference推理引擎的介绍，可以参考[Paddle Inference官网教程](https://www.paddlepaddle.org.cn/documentation/docs/zh/guides/infer/inference/inference_cn.html)。
+
+Inference 的获取可以参考 [ResNet50 推理模型准备](./ResNet.md#4.1) 。
+
+<a name="4.2"></a>
+
+### 4.2 基于 Python 预测引擎推理
+
+PaddleClas 提供了基于 python 预测引擎推理的示例。您可以参考[ResNet50 基于 Python 预测引擎推理](./ResNet.md#42-基于-python-预测引擎推理) 。
+
+<a name="4.3"></a>
+
+### 4.3 基于 C++ 预测引擎推理
+
+PaddleClas 提供了基于 C++ 预测引擎推理的示例，您可以参考[服务器端 C++ 预测](../../deployment/image_classification/cpp/linux.md)来完成相应的推理部署。如果您使用的是 Windows 平台，可以参考[基于 Visual Studio 2019 Community CMake 编译指南](../../deployment/image_classification/cpp/windows.md)完成相应的预测库编译和模型预测工作。
+
+<a name="4.4"></a>
+
+### 4.4 服务化部署
+
+Paddle Serving 提供高性能、灵活易用的工业级在线推理服务。Paddle Serving 支持 RESTful、gRPC、bRPC 等多种协议，提供多种异构硬件和多种操作系统环境下推理解决方案。更多关于Paddle Serving 的介绍，可以参考[Paddle Serving 代码仓库](https://github.com/PaddlePaddle/Serving)。
+
+PaddleClas 提供了基于 Paddle Serving 来完成模型服务化部署的示例，您可以参考[模型服务化部署](../../deployment/image_classification/paddle_serving.md)来完成相应的部署工作。
+
+<a name="4.5"></a>
+
+### 4.5 端侧部署
+
+Paddle Lite 是一个高性能、轻量级、灵活性强且易于扩展的深度学习推理框架，定位于支持包括移动端、嵌入式以及服务器端在内的多硬件平台。更多关于 Paddle Lite 的介绍，可以参考[Paddle Lite 代码仓库](https://github.com/PaddlePaddle/Paddle-Lite)。
+
+PaddleClas 提供了基于 Paddle Lite 来完成模型端侧部署的示例，您可以参考[端侧部署](../../deployment/image_classification/paddle_lite.md)来完成相应的部署工作。
+
+<a name="4.6"></a>
+
+### 4.6 Paddle2ONNX 模型转换与预测
+
+Paddle2ONNX 支持将 PaddlePaddle 模型格式转化到 ONNX 模型格式。通过 ONNX 可以完成将 Paddle 模型到多种推理引擎的部署，包括TensorRT/OpenVINO/MNN/TNN/NCNN，以及其它对 ONNX 开源格式进行支持的推理引擎或硬件。更多关于 Paddle2ONNX 的介绍，可以参考[Paddle2ONNX 代码仓库](https://github.com/PaddlePaddle/Paddle2ONNX)。
+
+PaddleClas 提供了基于 Paddle2ONNX 来完成 inference 模型转换 ONNX 模型并作推理预测的示例，您可以参考[Paddle2ONNX 模型转换与预测](../../deployment/image_classification/paddle2onnx.md)来完成相应的部署工作。

ppcls/arch/backbone/__init__.py

@@ -54,7 +54,7 @@ from .model_zoo.regnet import RegNetX_200MF, RegNetX_400MF, RegNetX_600MF, RegNe
 from .model_zoo.vision_transformer import ViT_small_patch16_224, ViT_base_patch16_224, ViT_base_patch16_384, ViT_base_patch32_384, ViT_large_patch16_224, ViT_large_patch16_384, ViT_large_patch32_384
 from .model_zoo.distilled_vision_transformer import DeiT_tiny_patch16_224, DeiT_small_patch16_224, DeiT_base_patch16_224, DeiT_tiny_distilled_patch16_224, DeiT_small_distilled_patch16_224, DeiT_base_distilled_patch16_224, DeiT_base_patch16_384, DeiT_base_distilled_patch16_384
 from .legendary_models.swin_transformer import SwinTransformer_tiny_patch4_window7_224, SwinTransformer_small_patch4_window7_224, SwinTransformer_base_patch4_window7_224, SwinTransformer_base_patch4_window12_384, SwinTransformer_large_patch4_window7_224, SwinTransformer_large_patch4_window12_384
-from .legendary_models.swin_transformerv2 import SwinTransformerV2_tiny_patch4_window8_256, SwinTransformerV2_small_patch4_window8_256, SwinTransformerV2_base_patch4_window8_256, SwinTransformerV2_tiny_patch4_window16_256, SwinTransformerV2_small_patch4_window16_256, SwinTransformerV2_base_patch4_window16_256, SwinTransformerV2_base_patch4_window12to16_256, SwinTransformerV2_base_patch4_window24_384, SwinTransformerV2_large_patch4_window16_256, SwinTransformerV2_large_patch4_window24_384
+from .model_zoo.swin_transformerv2 import SwinTransformerV2_tiny_patch4_window8_256, SwinTransformerV2_small_patch4_window8_256, SwinTransformerV2_base_patch4_window8_256, SwinTransformerV2_tiny_patch4_window16_256, SwinTransformerV2_small_patch4_window16_256, SwinTransformerV2_base_patch4_window16_256, SwinTransformerV2_base_patch4_window12to16_256, SwinTransformerV2_base_patch4_window24_384, SwinTransformerV2_large_patch4_window16_256, SwinTransformerV2_large_patch4_window24_384
 from .model_zoo.cswin_transformer import CSWinTransformer_tiny_224, CSWinTransformer_small_224, CSWinTransformer_base_224, CSWinTransformer_large_224, CSWinTransformer_base_384, CSWinTransformer_large_384
 from .model_zoo.mixnet import MixNet_S, MixNet_M, MixNet_L
 from .model_zoo.rexnet import ReXNet_1_0, ReXNet_1_3, ReXNet_1_5, ReXNet_2_0, ReXNet_3_0

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window12to16_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window16_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window24_384.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval
@@ -123,7 +123,7 @@ DataLoader:
             order: ''
     sampler:
       name: DistributedBatchSampler
-      batch_size: 128
+      batch_size: 32
       drop_last: False
       shuffle: False
     loader:
@@ -138,7 +138,7 @@ Infer:
         to_rgb: True
         channel_first: False
     - ResizeImage:
-        resize_short: 438
+        resize_short: 384
         interpolation: bicubic
         backend: pil
     - CropImage:

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window8_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_large_patch4_window16_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_large_patch4_window24_384.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval
@@ -94,7 +94,7 @@ DataLoader:
 
     sampler:
       name: DistributedBatchSampler
-      batch_size: 128
+      batch_size: 16
       drop_last: False
       shuffle: True
     loader:
@@ -123,7 +123,7 @@ DataLoader:
             order: ''
     sampler:
       name: DistributedBatchSampler
-      batch_size: 128
+      batch_size: 16
       drop_last: False
       shuffle: False
     loader:
@@ -138,7 +138,7 @@ Infer:
         to_rgb: True
         channel_first: False
     - ResizeImage:
-        resize_short: 438
+        resize_short: 384
         interpolation: bicubic
         backend: pil
     - CropImage:

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_small_patch4_window16_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_small_patch4_window8_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window16_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window8_256.yaml

@@ -44,9 +44,9 @@ Optimizer:
     # for 8 cards
     name: Cosine
     learning_rate: 1e-3
-    eta_min: 2e-5
+    eta_min: 1e-5
     warmup_epoch: 20
-    warmup_start_lr: 2e-6
+    warmup_start_lr: 1e-6
 
 
 # data loader for train and eval

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_base_patch4_window12to16_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_base_patch4_window12to16_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_base_patch4_window12to16_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window12to16_256.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/SwinTransformerV2/SwinTransformerV2_base_patch4_window12to16_256.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/SwinTransformerV2_base_patch4_window12to16_256_22kto1k_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_base_patch4_window16_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_base_patch4_window16_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_base_patch4_window16_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window16_256.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/SwinTransformerV2/SwinTransformerV2_base_patch4_window16_256.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/SwinTransformerV2_base_patch4_window16_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_base_patch4_window24_384_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_base_patch4_window24_384
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_base_patch4_window24_384.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window24_384.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/SwinTransformerV2/SwinTransformerV2_base_patch4_window24_384.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/SwinTransformerV2_base_patch4_window24_384_22kto1k_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,384,384]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_base_patch4_window8_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_base_patch4_window8_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_base_patch4_window8_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_base_patch4_window8_256.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/SwinTransformerV2/SwinTransformerV2_base_patch4_window8_256.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/SwinTransformerV2_base_patch4_window8_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_large_patch4_window16_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_large_patch4_window16_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_large_patch4_window16_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_large_patch4_window16_256.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/SwinTransformerV2/SwinTransformerV2_large_patch4_window16_256.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/SwinTransformerV2_large_patch4_window16_256_22kto1k_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_large_patch4_window24_384_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_large_patch4_window24_384
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_large_patch4_window24_384.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_large_patch4_window24_384.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/SwinTransformerV2/SwinTransformerV2_large_patch4_window24_384.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/SwinTransformerV2_large_patch4_window24_384_22kto1k_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,384,384]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_small_patch4_window16_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_small_patch4_window16_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_small_patch4_window16_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_small_patch4_window16_256.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/SwinTransformerV2/SwinTransformerV2_small_patch4_window16_256.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/SwinTransformerV2_small_patch4_window16_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_small_patch4_window8_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_small_patch4_window8_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_small_patch4_window8_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_small_patch4_window8_256.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/SwinTransformerV2/SwinTransformerV2_small_patch4_window8_256.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/SwinTransformerV2_small_patch4_window8_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window16_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_tiny_patch4_window16_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window16_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window16_256.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/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window16_256.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/SwinTransformerV2_tiny_patch4_window16_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file

test_tipc/configs/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window8_256_train_infer_python.txt

+===========================train_params===========================
+model_name:SwinTransformerV2_tiny_patch4_window8_256
+python:python3.7
+gpu_list:0|0,1
+-o Global.device:gpu
+-o Global.auto_cast:null
+-o Global.epochs:lite_train_lite_infer=2|whole_train_whole_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/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window8_256.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 -o Global.eval_during_train=False -o Global.save_interval=2
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:tools/eval.py -c ppcls/configs/ImageNet/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window8_256.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/SwinTransformerV2/SwinTransformerV2_tiny_patch4_window8_256.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/SwinTransformerV2_tiny_patch4_window8_256_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:False
+-o Global.cpu_num_threads:1
+-o Global.batch_size:1
+-o Global.use_tensorrt:False
+-o Global.use_fp16:False
+-o Global.inference_model_dir:../inference
+-o Global.infer_imgs:../dataset/ILSVRC2012/val/ILSVRC2012_val_00000001.JPEG
+-o Global.save_log_path:null
+-o Global.benchmark:False
+null:null
+null:null
+===========================train_benchmark_params==========================
+batch_size:104|128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,256,256]}]
\ No newline at end of file