diff --git a/README_cn.md b/README_cn.md index ebc2dfa3c701bbf0ba4952c2bee44ef135b600ef..083cc3f22d01871ad74ae18aa5478adca56ae9fc 100644 --- a/README_cn.md +++ b/README_cn.md @@ -33,7 +33,7 @@ - 发布行人分析工具[PP-Human v2](./deploy/pipeline)，新增打架、打电话、抽烟、闯入四大行为识别，底层算法性能升级，覆盖行人检测、跟踪、属性三类核心算法能力，提供保姆级全流程开发及模型优化策略，支持在线视频流输入 - 首次发布[PP-Vehicle](./deploy/pipeline)，提供车牌识别、车辆属性分析（颜色、车型）、车流量统计以及违章检测四大功能，兼容图片、在线视频流、视频输入，提供完善的二次开发文档教程 - 💡 前沿算法： - - 全面覆盖的[YOLO家族](docs/feature_models/YOLOSERIES_MODEL.md)经典与最新模型代码库[PaddleDetection_YOLOSeries](https://github.com/nemonameless/PaddleDetection_YOLOSeries): 包括YOLOv3，百度飞桨自研的实时高精度目标检测模型PP-YOLOE，以及前沿检测算法YOLOv4、YOLOv5、YOLOX，MT-YOLOv6及YOLOv7 + - 全面覆盖的[YOLO家族](docs/feature_models/YOLOSERIES_MODEL.md)经典与最新模型代码库[PaddleDetection_YOLOSeries](https://github.com/nemonameless/PaddleDetection_YOLOSeries): 包括YOLOv3，百度飞桨自研的实时高精度目标检测模型PP-YOLOE，以及前沿检测算法YOLOv4、YOLOv5、YOLOX，YOLOv6及YOLOv7 - 新增基于[ViT](configs/vitdet)骨干网络高精度检测模型，COCO数据集精度达到55.7% mAP；新增[OC-SORT](configs/mot/ocsort)多目标跟踪模型；新增[ConvNeXt](configs/convnext)骨干网络 - 📋 产业范例：新增[智能健身](https://aistudio.baidu.com/aistudio/projectdetail/4385813)、[打架识别](https://aistudio.baidu.com/aistudio/projectdetail/4086987?channelType=0&channel=0)、[来客分析](https://aistudio.baidu.com/aistudio/projectdetail/4230123?channelType=0&channel=0)、车辆结构化范例 @@ -124,7 +124,7 @@
  • RetinaNet
  • YOLOv3
  • YOLOv5
    • -
  • MT-YOLOv6
    • +
  • YOLOv6
  • YOLOv7
  • PP-YOLOv1/v2
  • PP-YOLO-Tiny
    • diff --git a/README_en.md b/README_en.md index 500397a4c1089d11aa81f4902cf49c8a485dabab..1dd35bb6254da77a428970ed5b2b974698e28fb2 100644 --- a/README_en.md +++ b/README_en.md @@ -38,7 +38,7 @@ - 💡 Cutting-edge algorithms： - - Covers [YOLO family](https://github.com/nemonameless/PaddleDetection_YOLOSeries) classic and latest models: YOLOv3, PP-YOLOE (a real-time high-precision object detection model developed by Baidu PaddlePaddle), and cutting-edge detection algorithms such as YOLOv4, YOLOv5, YOLOX, MT-YOLOv6, and YOLOv7 + - Covers [YOLO family](https://github.com/nemonameless/PaddleDetection_YOLOSeries) classic and latest models: YOLOv3, PP-YOLOE (a real-time high-precision object detection model developed by Baidu PaddlePaddle), and cutting-edge detection algorithms such as YOLOv4, YOLOv5, YOLOX, YOLOv6, and YOLOv7 - Newly add high precision detection model based on [ViT](configs/vitdet) backbone network, with a 55.7% mAP accuracy on COCO dataset; newly add multi-object tracking model [OC-SORT](configs/mot/ocsort); newly add [ConvNeXt](configs/convnext) backbone network. - 📋 Industrial applications: Newly add [Smart Fitness](https://aistudio.baidu.com/aistudio/projectdetail/4385813), [Fighting recognition](https://aistudio.baidu.com/aistudio/projectdetail/4086987?channelType=0&channel=0),[ and Visitor Analysis](https://aistudio.baidu.com/aistudio/projectdetail/4230123?channelType=0&channel=0). diff --git a/docs/CHANGELOG.md b/docs/CHANGELOG.md index f5fef04f1c946f69bda11ee16d7222c0dc6e8eab..86b1bf0f7463e08861a95c770d49d2c89785953f 100644 --- a/docs/CHANGELOG.md +++ b/docs/CHANGELOG.md @@ -35,7 +35,7 @@ - 前沿算法 - YOLO家族全系列模型 - - 发布YOLO家族全系列模型，覆盖前沿检测算法YOLOv5、MT-YOLOv6及YOLOv7 + - 发布YOLO家族全系列模型，覆盖前沿检测算法YOLOv5、YOLOv6及YOLOv7 - 基于ConvNext骨干网络，YOLO各算法训练周期缩5-8倍，精度普遍提升1%-5% mAP；使用模型压缩策略实现精度无损的同时速度提升30%以上 - 新增基于ViT骨干网络高精度检测模型，COCO数据集精度达到55.7% mAP - 新增OC-SORT多目标跟踪模型 diff --git a/docs/CHANGELOG_en.md b/docs/CHANGELOG_en.md index 601bb2c5b027bb100eb55aef34aebf221cd555dc..15b8321e941ef0e4e39dd050e1c96afe09b96356 100644 --- a/docs/CHANGELOG_en.md +++ b/docs/CHANGELOG_en.md @@ -38,7 +38,7 @@ English | [简体中文](./CHANGELOG.md) - Cutting-edge algorithms - YOLO Family - - Release the full range of YOLO family models covering the cutting-edge detection algorithms YOLOv5, MT-YOLOv6 and YOLOv7 + - Release the full range of YOLO family models covering the cutting-edge detection algorithms YOLOv5, YOLOv6 and YOLOv7 - Based on the ConvNext backbone network, YOLO's algorithm training periods are reduced by 5-8 times with accuracy generally improving by 1%-5% mAP; Thanks to the model compression strategy, its speed increased by over 30% with no loss of precision. - Newly add high precision detection model based on [ViT](configs/vitdet) backbone network, with a 55.7% mAP accuracy on the COCO dataset - Newly add multi-object tracking model [OC-SORT](configs/mot/ocsort) diff --git a/docs/MODEL_ZOO_cn.md b/docs/MODEL_ZOO_cn.md index 7e3fa492504e567f5c8b6d0f01cc52dace27daa6..a98d24e4c924e8525b38b696b315bdfa4c585179 100644 --- a/docs/MODEL_ZOO_cn.md +++ b/docs/MODEL_ZOO_cn.md @@ -93,7 +93,19 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型 ### YOLOX -请参考[YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/yolox) +请参考[YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolox) + +### YOLOv5 + +请参考[YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5) + +### YOLOv6 + +请参考[YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6) + +### YOLOv7 + +请参考[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7) ## 旋转框检测 diff --git a/docs/MODEL_ZOO_en.md b/docs/MODEL_ZOO_en.md index 31b3a64969b175a9d9ad557d22f29735ec665aa1..e201287a78eb4dbb4e8b1eb190366c2bfa8e30f1 100644 --- a/docs/MODEL_ZOO_en.md +++ b/docs/MODEL_ZOO_en.md @@ -92,7 +92,19 @@ Please refer to[PP-YOLOE](https://github.com/PaddlePaddle/PaddleDetection/tree/r ### YOLOX -Please refer to[YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/yolox) +Please refer to[YOLOX](https://github.com/PaddlePaddle/PaddleDetection/tree/develop/configs/yolox) + +### YOLOv5 + +Please refer to[YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5) + +### YOLOv6 + +Please refer to[YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6) + +### YOLOv7 + +Please refer to[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7) ## Rotating frame detection diff --git a/docs/feature_models/YOLOSERIES_MODEL.md b/docs/feature_models/YOLOSERIES_MODEL.md index 52c4528a52642af5be27fc00cfe5c13b107d73aa..27c00be8793f4457d4ec11b44519f02112815c0a 100644 --- a/docs/feature_models/YOLOSERIES_MODEL.md +++ b/docs/feature_models/YOLOSERIES_MODEL.md @@ -1,6 +1,6 @@ 简体中文 | [English](YOLOSERIES_MODEL_en.md) -# YOLOSeries +# [**YOLOSeries**](https://github.com/nemonameless/PaddleDetection_YOLOSeries) ## 内容 - [简介](#简介) @@ -8,7 +8,7 @@ - [PP-YOLOE](#PP-YOLOE) - [YOLOX](#YOLOX) - [YOLOv5](#YOLOv5) - - [MT-YOLOv6](#MT-YOLOv6) + - [YOLOv6](#YOLOv6) - [YOLOv7](#YOLOv7) - [使用指南](#使用指南) - [一键运行全流程](#一键运行全流程) @@ -16,20 +16,41 @@ ## 简介 -[**YOLOSeries**](https://github.com/nemonameless/PaddleDetection_YOLOSeries)是基于[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)的YOLO系列模型库，**由PaddleDetection团队成员建设和维护**，支持`YOLOv3`,`PP-YOLOE`,`PP-YOLOE+`,`YOLOX`,`YOLOv5`,`MT-YOLOv6`,`YOLOv7`等模型，其upstream为PaddleDetection的[develop](https://github.com/PaddlePaddle/PaddleDetection/tree/develop)分支，并与PaddleDetection主代码库分支保持同步更新，包括github和gitee的代码，欢迎一起使用和建设！ +[**YOLOSeries**](https://github.com/nemonameless/PaddleDetection_YOLOSeries)是基于[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)的YOLO系列模型库，**由PaddleDetection团队成员建设和维护**，支持`YOLOv3`,`PP-YOLOE`,`PP-YOLOE+`,`YOLOX`,`YOLOv5`,`YOLOv6`,`YOLOv7`等模型，其upstream为PaddleDetection的[develop](https://github.com/PaddlePaddle/PaddleDetection/tree/develop)分支，并与PaddleDetection主代码库分支保持同步更新，包括github和gitee的代码，欢迎一起使用和建设！ + +## Updates! +* 【2022/09/21】精简代码库只保留主要的YOLO模型相关的代码(release/2.5 branch)； +* 【2022/09/19】支持[`YOLOv6`](configs/yolov6)新版，包括n/t/s/m/l模型； +* 【2022/08/23】发布`PaddleDetection_YOLOSeries`代码库: 支持`YOLOv3`,`PP-YOLOE`,`PP-YOLOE+`,`YOLOX`,`YOLOv5`,`MT-YOLOv6`,`YOLOv7`等YOLO模型，支持ConvNeXt骨干网络高精度版`PP-YOLOE`,`YOLOX`和`YOLOv5`等模型，支持PaddleSlim无损加速量化训练`PP-YOLOE`,`YOLOv5`,`MT-YOLOv6`和`YOLOv7`等模型，详情可阅读[此文章](https://mp.weixin.qq.com/s/Hki01Zs2lQgvLSLWS0btrA)； + **注意:** + - 此代码库**推荐使用paddlepaddle-2.3.0以上的版本**，请参考[官网](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/install/pip/linux-pip.html)下载对应适合版本，**其中develop分支代码请安装paddle develop版本，其余分支建议安装paddle 2.3.2版本**。 - github链接为：https://github.com/nemonameless/PaddleDetection_YOLOSeries - gitee链接为：https://gitee.com/nemonameless/PaddleDetection_YOLOSeries - 提issue可以在此代码库的[issues](https://github.com/nemonameless/PaddleDetection_YOLOSeries/issues)页面中，也可以在[PaddleDetection issues](https://github.com/PaddlePaddle/PaddleDetection/issues)中，也欢迎提[PR](https://github.com/nemonameless/PaddleDetection_YOLOSeries/pulls)共同建设和维护。 - - [PP-YOLOE](../../configs/ppyoloe),[PP-YOLOE+](../../configs/ppyoloe),[PP-YOLO](../../configs/ppyolo),[PP-YOLOv2](../../configs/ppyolo),[YOLOv3](../../configs/yolov3)和[YOLOX](../../configs/yolox)等模型推荐在[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)中使用，**会最先发布PP-YOLO系列特色检测模型的最新进展**。 - - [YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5),[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7)和[MT-YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt)模型推荐在此代码库中使用，**由于GPL开源协议而不合入[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)主代码库**。 - - `YOLOSeries`代码库**推荐使用paddlepaddle-2.3.0及以上的版本**，请参考[官网](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/install/pip/linux-pip.html)下载对应适合版本。 + - [PP-YOLOE](configs/ppyoloe),[PP-YOLOE+](configs/ppyoloe),[PP-YOLO](configs/ppyolo),[PP-YOLOv2](configs/ppyolo),[YOLOv3](configs/yolov3)和[YOLOX](configs/yolox)等模型推荐在[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)中使用，**会最先发布PP-YOLO系列特色检测模型的最新进展**。 + - [YOLOv5](configs/yolov5),[YOLOv7](configs/yolov7)和[YOLOv6](configs/yolov6)模型推荐在此代码库中使用，**由于GPL开源协议而不合入[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)主代码库**。 + + +## 技术交流 + +- 如果你发现任何PaddleDetection存在的问题或者是建议, 欢迎通过[GitHub Issues](https://github.com/PaddlePaddle/PaddleDetection/issues)给我们提issues。 + +- **欢迎加入PaddleDetection 微信用户群（扫码填写问卷即可入群）** + - **入群福利 💎：获取PaddleDetection团队整理的重磅学习大礼包🎁** + - 📊 福利一：获取飞桨联合业界企业整理的开源数据集 + - 👨‍🏫 福利二：获取PaddleDetection历次发版直播视频与最新直播咨询 + - 🗳 福利三：获取垂类场景预训练模型集合，包括工业、安防、交通等5+行业场景 + - 🗂 福利四：获取10+全流程产业实操范例，覆盖火灾烟雾检测、人流量计数等产业高频场景 +
    + +
    ## 模型库 -### [PP-YOLOE](../../configs/ppyoloe) +### [PP-YOLOE, PP-YOLOE+](configs/ppyoloe) | 网络模型 | 输入尺寸 | 图片数/GPU | 学习率策略 | 推理耗时(ms) | mAPval
    0.5:0.95 | mAPval
    0.5 | Params(M) | FLOPs(G) | 下载链接 | 配置文件 | | :------------- | :------- | :-------: | :------: | :------------: | :---------------------: | :----------------: |:---------: | :------: |:---------------: |:-----: | @@ -106,23 +127,32 @@ | YOLOv5-l | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_l_300e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_l_300e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_l_300e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_l_300e_coco_wo_nms.onnx) | | YOLOv5-x | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_x_300e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_x_300e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_x_300e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov5/yolov5_x_300e_coco_wo_nms.onnx) | -### [MT-YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt) -| 网络模型 | 输入尺寸 | 图片数/GPU | 学习率策略 | 推理耗时(ms) | mAPval
    0.5:0.95 | mAPval
    0.5 | Params(M) | FLOPs(G) | 下载链接 | 配置文件 | +### [YOLOv6](configs/yolov6) + +| 网络网络 | 输入尺寸 | 图片数/GPU | 学习率策略 | 模型推理耗时(ms) | mAP | AP50 | Params(M) | FLOPs(G) | 下载链接 | 配置文件 | | :------------- | :------- | :-------: | :------: | :---------: | :-----: |:-----: | :-----: |:-----: | :-------------: | :-----: | -| *YOLOv6mt-n | 416 | 32 | 400e | 2.5 | 30.5 | 46.8 | 4.74 | 5.16 |[model](https://paddledet.bj.bcebos.com/models/yolov6mt_n_416_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt/yolov6mt_n_416_400e_coco.yml) | -| *YOLOv6mt-n | 640 | 32 | 400e | 2.8 | 34.7 | 52.7 | 4.74 | 12.2 |[model](https://paddledet.bj.bcebos.com/models/yolov6mt_n_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt/yolov6mt_n_400e_coco.yml) | -| *YOLOv6mt-t | 640 | 32 | 400e | 2.9 | 40.8 | 60.4 | 16.36 | 39.94 |[model](https://paddledet.bj.bcebos.com/models/yolov6mt_t_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt/yolov6mt_t_400e_coco.yml) | -| *YOLOv6mt-s | 640 | 32 | 400e | 3.0 | 42.5 | 61.7 | 18.87 | 48.36 |[model](https://paddledet.bj.bcebos.com/models/yolov6mt_s_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt/yolov6mt_s_400e_coco.yml) | +| *YOLOv6-n | 416 | 32 | 400e | 1.0 | 31.1 | 45.3 | 4.74 | 5.16 |[model](https://paddledet.bj.bcebos.com/models/yolov6_n_416_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_n_416_400e_coco.yml) | +| *YOLOv6-n | 640 | 32 | 400e | 1.3 | 36.1 | 51.9 | 4.74 | 12.21 |[model](https://paddledet.bj.bcebos.com/models/yolov6_n_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_n_400e_coco.yml) | +| *YOLOv6-t | 640 | 32 | 400e | 2.1 | 40.7 | 57.4 | 10.63 | 27.29 |[model](https://paddledet.bj.bcebos.com/models/yolov6_t_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_t_400e_coco.yml) | +| *YOLOv6-s | 640 | 32 | 400e | 2.6 | 43.4 | 60.5 | 18.87 | 48.35 |[model](https://paddledet.bj.bcebos.com/models/yolov6_s_400e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_s_400e_coco.yml) | +| *YOLOv6-m | 640 | 32 | 300e | 5.0 | 49.0 | 66.5 | 37.17 | 88.82 |[model](https://paddledet.bj.bcebos.com/models/yolov6_m_300e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_m_300e_coco.yml) | +| *YOLOv6-l | 640 | 32 | 300e | 7.9 | 51.0 | 68.9 | 63.54 | 155.89 |[model](https://paddledet.bj.bcebos.com/models/yolov6_l_300e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_l_300e_coco.yml) | +| *YOLOv6-l-silu | 640 | 32 | 300e | 9.6 | 51.7 | 69.6 | 58.59 | 142.66 |[model](https://paddledet.bj.bcebos.com/models/yolov6_l_silu_300e_coco.pdparams) | [config](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6/yolov6_l_silu_300e_coco.yml) | + #### 部署模型 | 网络模型 | 输入尺寸 | 导出后的权重(w/o NMS) | ONNX(w/o NMS) | | :-------- | :--------: | :---------------------: | :----------------: | -| YOLOv6mt-n | 416 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_416_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_416_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_416_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_416_400e_coco_wo_nms.onnx) | -| YOLOv6mt-n | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_n_400e_coco_wo_nms.onnx) | -| YOLOv6mt-t | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_t_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_t_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_t_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_t_400e_coco_wo_nms.onnx) | -| YOLOv6mt-s | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_s_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_s_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_s_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6mt/yolov6mt_s_400e_coco_wo_nms.onnx) | +| yolov6-n | 416 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_416_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_416_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_416_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_416_400e_coco_wo_nms.onnx) | +| yolov6-n | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_n_400e_coco_wo_nms.onnx) | +| yolov6-t | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_t_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_t_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_t_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_t_400e_coco_wo_nms.onnx) | +| yolov6-s | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_s_400e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_s_400e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_s_400e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_s_400e_coco_wo_nms.onnx) | +| yolov6-m | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_m_300e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_m_300e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_m_300e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_m_300e_coco_wo_nms.onnx) | +| yolov6-l | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_300e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_300e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_300e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_300e_coco_wo_nms.onnx) | +| yolov6-l-silu | 640 | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_silu_300e_coco_w_nms.zip) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_silu_300e_coco_wo_nms.zip) | [( w/ nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_silu_300e_coco_w_nms.onnx) | [( w/o nms)](https://paddledet.bj.bcebos.com/deploy/yoloseries/yolov6/yolov6_l_silu_300e_coco_wo_nms.onnx) | + ### [YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7) @@ -156,7 +186,7 @@ ### **注意:** - 所有模型均使用COCO train2017作为训练集，在COCO val2017上验证精度，模型前带*表示训练更新中。 - - 具体精度和速度细节请查看[PP-YOLOE](../../configs/ppyoloe),[YOLOX](../../configs/yolox),[YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5),[MT-YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt),[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7)。 + - 具体精度和速度细节请查看[PP-YOLOE](../../configs/ppyoloe),[YOLOX](../../configs/yolox),[YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5),[YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs),[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7)。 - 模型推理耗时(ms)为TensorRT-FP16下测试的耗时，不包含数据预处理和模型输出后处理(NMS)的耗时。测试采用单卡V100，batch size=1，测试环境为**paddlepaddle-2.3.0**, **CUDA 11.2**, **CUDNN 8.2**, **GCC-8.2**, **TensorRT 8.0.3.4**，具体请参考各自模型主页。 - **统计参数量Params(M)**，可以将以下代码插入[trainer.py](https://github.com/nemonameless/PaddleDetection_YOLOSeries/blob/develop/ppdet/engine/trainer.py#L150)。 ```python @@ -180,7 +210,7 @@ 并重新导出，使用时再**另接自己写的NMS后处理**。 - 基于[PaddleSlim](https://github.com/PaddlePaddle/PaddleSlim)对YOLO系列模型进行量化训练，可以实现精度基本无损，速度普遍提升30%以上，具体请参照[模型自动化压缩工具ACT](https://github.com/PaddlePaddle/PaddleSlim/tree/develop/example/auto_compression)。 - [PP-YOLOE](../../configs/ppyoloe),[PP-YOLOE+](../../configs/ppyoloe),[YOLOv3](../../configs/yolov3)和[YOLOX](../../configs/yolox)推荐在[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)里使用，会最先发布**PP-YOLO系列特色检测模型的最新进展**。 - - [YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5),[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7)和[MT-YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6mt)由于GPL协议而不合入[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)主代码库。 + - [YOLOv5](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov5),[YOLOv7](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov7)和[YOLOv6](https://github.com/nemonameless/PaddleDetection_YOLOSeries/tree/develop/configs/yolov6)由于GPL协议而不合入[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)主代码库。 - **paddlepaddle版本推荐使用2.3.0版本以上**。