diff --git a/GETTING_STARTED.md b/GETTING_STARTED.md index 363aa1f9..0894d2d4 100644 --- a/GETTING_STARTED.md +++ b/GETTING_STARTED.md @@ -5,9 +5,9 @@ This document provides a brief introduction to the usage of built-in command-lin ### Inference Demo with Pre-trained Models 1. Pick a model and its config file from the - [model zoo](MODEL_ZOO.md), + [Model Zoo](benchmark_results.md), such as, `./configs/yolov7/yolov7.yaml`. -2. Download the corresponding pre-trained checkpoint from the [model zoo](MODEL_ZOO.md) of each model. +2. Download the corresponding pre-trained checkpoint from the [Model Zoo](benchmark_results.md) of each model. 3. To run YOLO object detection with the built-in configs, please run: ``` diff --git a/GETTING_STARTED_CN.md b/GETTING_STARTED_CN.md index 807e5d6e..5daeb397 100644 --- a/GETTING_STARTED_CN.md +++ b/GETTING_STARTED_CN.md @@ -4,8 +4,8 @@ ### 使用预训练模型进行推理 -1. 从[model zoo](MODEL_ZOO.md)中选择一个模型及其配置文件，例如， `./configs/yolov7/yolov7.yaml`. -2. 从[model zoo](MODEL_ZOO.md)中下载相应的预训练模型权重文件。 +1. 从[模型仓库](benchmark_results.md)中选择一个模型及其配置文件，例如， `./configs/yolov7/yolov7.yaml`. +2. 从[模型仓库](benchmark_results.md)中下载相应的预训练模型权重文件。 3. 使用内置配置进行推理，请运行以下命令： ```shell diff --git a/MODEL_ZOO.md b/MODEL_ZOO.md deleted file mode 100644 index 4e6cc49a..00000000 --- a/MODEL_ZOO.md +++ /dev/null @@ -1,43 +0,0 @@ -# MindYOLO Model Zoo and Baselines - -## Detection - -| Name | Scale | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|:--------------------:|----------|-----------|--------------|-------------|--------|--------|--------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------| -| YOLOv8 | N | D910x8-G | 640 | MS COCO 2017 | 37.2 | 3.2M | 8.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-cc07f5bd.ckpt) | -| YOLOv8 | S | D910x8-G | 640 | MS COCO 2017 | 44.6 | 11.2M | 28.6G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-3086f0c9.ckpt) | -| YOLOv8 | M | D910x8-G | 640 | MS COCO 2017 | 50.5 | 25.9M | 78.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-m_500e_mAP505-8ff7a728.ckpt) | -| YOLOv8 | L | D910x8-G | 640 | MS COCO 2017 | 52.8 | 43.7M | 165.2G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-l_500e_mAP528-6e96d6bb.ckpt) | -| YOLOv8 | X | D910x8-G | 640 | MS COCO 2017 | 53.7 | 68.2M | 257.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x_500e_mAP537-b958e1c7.ckpt) | -| YOLOv7 | Tiny | D910x8-G | 640 | MS COCO 2017 | 37.5 | 6.2M | 13.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-d8972c94.ckpt) | -| YOLOv7 | L | D910x8-G | 640 | MS COCO 2017 | 50.8 | 36.9M | 104.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7_300e_mAP508-734ac919.ckpt) | -| YOLOv7 | X | D910x8-G | 640 | MS COCO 2017 | 52.4 | 71.3M | 189.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-x_300e_mAP524-e2f58741.ckpt) | -| YOLOv5 | N | D910x8-G | 640 | MS COCO 2017 | 27.3 | 1.9M | 4.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-9b16bd7b.ckpt) | -| YOLOv5 | S | D910x8-G | 640 | MS COCO 2017 | 37.6 | 7.2M | 16.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-860bcf3b.ckpt) | -| YOLOv5 | M | D910x8-G | 640 | MS COCO 2017 | 44.9 | 21.2M | 49.0G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5m_300e_mAP449-e7bbf695.ckpt) | -| YOLOv5 | L | D910x8-G | 640 | MS COCO 2017 | 48.5 | 46.5M | 109.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5l_300e_mAP485-a28bce73.ckpt) | -| YOLOv5 | X | D910x8-G | 640 | MS COCO 2017 | 50.5 | 86.7M | 205.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5x_300e_mAP505-97d36ddc.ckpt) | -| YOLOv4 | CSPDarknet53 | D910x8-G | 608 | MS COCO 2017 | 45.4 | 27.6M | 52G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov4/yolov4.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-50172f93.ckpt) | -| YOLOv4 | CSPDarknet53(silu) | D910x8-G | 608 | MS COCO 2017 | 45.8 | 27.6M | 52G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov4/yolov4-silu.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_silu_320e_map458-bdfc3205.ckpt) | -| YOLOv3 | Darknet53 | D910x8-G | 640 | MS COCO 2017 | 45.5 | 61.9M | 156.4G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov3/yolov3.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-adfb27af.ckpt) | -| YOLOX | N | D910x8-G | 416 | MS COCO 2017 | 24.1 | 0.9M | 1.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-nano.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-n_300e_map241-ec9815e3.ckpt) | -| YOLOX | Tiny | D910x8-G | 416 | MS COCO 2017 | 33.3 | 5.1M | 6.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-tiny_300e_map333-e5ae3a2e.ckpt) | -| YOLOX | S | D910x8-G | 640 | MS COCO 2017 | 40.7 | 9.0M | 26.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-0983e07f.ckpt) | -| YOLOX | M | D910x8-G | 640 | MS COCO 2017 | 46.7 | 25.3M | 73.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-m_300e_map467-1db321ee.ckpt) | -| YOLOX | L | D910x8-G | 640 | MS COCO 2017 | 49.2 | 54.2M | 155.6G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-l_300e_map492-52a4ab80.ckpt) | -| YOLOX | X | D910x8-G | 640 | MS COCO 2017 | 51.6 | 99.1M | 281.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-x_300e_map516-52216d90.ckpt) | -| YOLOX | Darknet53 | D910x8-G | 640 | MS COCO 2017 | 47.7 | 63.7M | 185.3G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-darknet53.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-darknet53_300e_map477-b5fcaba9.ckpt) | - -## Segmentation - -| Name | Scale | Context | ImageSize | Dataset | Box mAP (%) | Mask mAP (%) | Params | FLOPs | Recipe | Download | -|------------|-------|----------|-----------|--------------|-------------|--------------|--------|--------|---------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------| -| YOLOv8-seg | X | D910x8-G | 640 | MS COCO 2017 | 52.5 | 42.9 | 71.8M | 344.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/seg/yolov8x-seg.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x-seg_300e_mAP_mask_429-b4920557.ckpt) | - -## Depoly inference - -- See [support list](./deploy/README.md) - -## Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. -- Box mAP: Accuracy reported on the validation set. diff --git a/README.md b/README.md index 1c010b0a..da23fb7c 100644 --- a/README.md +++ b/README.md @@ -14,16 +14,20 @@ MindYOLO implements state-of-the-art YOLO series algorithms based on MindSpore. The following is the corresponding `mindyolo` versions and supported `mindspore` versions. -| mindyolo | mindspore | -| :--: | :--: | -| master | master | -| 0.4 | 2.3.0 | -| 0.3 | 2.2.10 | -| 0.2 | 2.0 | -| 0.1 | 1.8 | +| mindyolo | mindspore | tested hardware | +| :--: | :--: | :--: | +| master | master | Ascend 910* | +| 0.4 | 2.3.0 | Ascend 910* | +| 0.3 | 2.2.10 | Ascend 910* & Ascend 910 | +| 0.2 | 2.0 | Ascend 910 | +| 0.1 | 1.8 | Ascend 910 | +## Benchmark and Model Zoo + +See [Benchmark Results](benchmark_results.md). + ## supported model list - [ ] YOLOv10 (welcome to contribute) - [ ] YOLOv9 (welcome to contribute) @@ -34,7 +38,6 @@ The following is the corresponding `mindyolo` versions and supported `mindspore` - [x] [YOLOv4](configs/yolov4) - [x] [YOLOv3](configs/yolov3) - ## Installation See [INSTALLATION](docs/en/installation.md) for details. diff --git a/benchmark_results.md b/benchmark_results.md new file mode 100644 index 00000000..bb5ddbca --- /dev/null +++ b/benchmark_results.md @@ -0,0 +1,63 @@ +# MindYOLO Benchmark and Baselines + +## Detection +
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv8 | N | 16 * 8 | 640 | MS COCO 2017 | 37.2 | 3.2M | [yaml](./configs/yolov8/yolov8n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-cc07f5bd.ckpt) | +| YOLOv8 | S | 16 * 8 | 640 | MS COCO 2017 | 44.6 | 11.2M | [yaml](./configs/yolov8/yolov8s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-3086f0c9.ckpt) | +| YOLOv8 | M | 16 * 8 | 640 | MS COCO 2017 | 50.5 | 25.9M | [yaml](./configs/yolov8/yolov8m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-m_500e_mAP505-8ff7a728.ckpt) | +| YOLOv8 | L | 16 * 8 | 640 | MS COCO 2017 | 52.8 | 43.7M | [yaml](./configs/yolov8/yolov8l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-l_500e_mAP528-6e96d6bb.ckpt) | +| YOLOv8 | X | 16 * 8 | 640 | MS COCO 2017 | 53.7 | 68.2M | [yaml](./configs/yolov8/yolov8x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x_500e_mAP537-b958e1c7.ckpt) | +| YOLOv7 | Tiny | 16 * 8 | 640 | MS COCO 2017 | 37.5 | 6.2M | [yaml](./configs/yolov7/yolov7-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-d8972c94.ckpt) | +| YOLOv7 | L | 16 * 8 | 640 | MS COCO 2017 | 50.8 | 36.9M | [yaml](./configs/yolov7/yolov7.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7_300e_mAP508-734ac919.ckpt) | +| YOLOv7 | X | 12 * 8 | 640 | MS COCO 2017 | 52.4 | 71.3M | [yaml](./configs/yolov7/yolov7-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-x_300e_mAP524-e2f58741.ckpt) | +| YOLOv5 | N | 32 * 8 | 640 | MS COCO 2017 | 27.3 | 1.9M | [yaml](./configs/yolov5/yolov5n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-9b16bd7b.ckpt) | +| YOLOv5 | S | 32 * 8 | 640 | MS COCO 2017 | 37.6 | 7.2M | [yaml](./configs/yolov5/yolov5s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-860bcf3b.ckpt) | +| YOLOv5 | M | 32 * 8 | 640 | MS COCO 2017 | 44.9 | 21.2M | [yaml](./configs/yolov5/yolov5m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5m_300e_mAP449-e7bbf695.ckpt) | +| YOLOv5 | L | 32 * 8 | 640 | MS COCO 2017 | 48.5 | 46.5M | [yaml](./configs/yolov5/yolov5l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5l_300e_mAP485-a28bce73.ckpt) | +| YOLOv5 | X | 16 * 8 | 640 | MS COCO 2017 | 50.5 | 86.7M | [yaml](./configs/yolov5/yolov5x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5x_300e_mAP505-97d36ddc.ckpt) | +| YOLOv4 | CSPDarknet53 | 16 * 8 | 608 | MS COCO 2017 | 45.4 | 27.6M | [yaml](./configs/yolov4/yolov4.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-50172f93.ckpt) | +| YOLOv4 | CSPDarknet53(silu) | 16 * 8 | 608 | MS COCO 2017 | 45.8 | 27.6M | [yaml](./configs/yolov4/yolov4-silu.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_silu_320e_map458-bdfc3205.ckpt) | +| YOLOv3 | Darknet53 | 16 * 8 | 640 | MS COCO 2017 | 45.5 | 61.9M | [yaml](./configs/yolov3/yolov3.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-adfb27af.ckpt) | +| YOLOX | N | 8 * 8 | 416 | MS COCO 2017 | 24.1 | 0.9M | [yaml](./configs/yolox/yolox-nano.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-n_300e_map241-ec9815e3.ckpt) | +| YOLOX | Tiny | 8 * 8 | 416 | MS COCO 2017 | 33.3 | 5.1M | [yaml](./configs/yolox/yolox-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-tiny_300e_map333-e5ae3a2e.ckpt) | +| YOLOX | S | 8 * 8 | 640 | MS COCO 2017 | 40.7 | 9.0M | [yaml](./configs/yolox/yolox-s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-0983e07f.ckpt) | +| YOLOX | M | 8 * 8 | 640 | MS COCO 2017 | 46.7 | 25.3M | [yaml](./configs/yolox/yolox-m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-m_300e_map467-1db321ee.ckpt) | +| YOLOX | L | 8 * 8 | 640 | MS COCO 2017 | 49.2 | 54.2M | [yaml](./configs/yolox/yolox-l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-l_300e_map492-52a4ab80.ckpt) | +| YOLOX | X | 8 * 8 | 640 | MS COCO 2017 | 51.6 | 99.1M | [yaml](./configs/yolox/yolox-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-x_300e_map516-52216d90.ckpt) | +| YOLOX | Darknet53 | 8 * 8 | 640 | MS COCO 2017 | 47.7 | 63.7M | [yaml](./configs/yolox/yolox-darknet53.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-darknet53_300e_map477-b5fcaba9.ckpt) | +
+ +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv8 | N | 16 * 8 | 640 | MS COCO 2017 | 37.3 | 373.55 | 3.2M | [yaml](./configs/yolov8/yolov8n.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-0e737186-910v2.ckpt) | +| YOLOv8 | S | 16 * 8 | 640 | MS COCO 2017 | 44.7 | 365.53 | 11.2M | [yaml](./configs/yolov8/yolov8s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-fae4983f-910v2.ckpt) | +| YOLOv7 | Tiny | 16 * 8 | 640 | MS COCO 2017 | 37.5 | 496.21 | 6.2M | [yaml](./configs/yolov7/yolov7-tiny.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-1d2ddf4b-910v2.ckpt) | +| YOLOv5 | N | 32 * 8 | 640 | MS COCO 2017 | 27.4 | 736.08 | 1.9M | [yaml](./configs/yolov5/yolov5n.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-bedf9a93-910v2.ckpt) | +| YOLOv5 | S | 32 * 8 | 640 | MS COCO 2017 | 37.6 | 787.34 | 7.2M | [yaml](./configs/yolov5/yolov5s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-df4a45b6-910v2.ckpt) | +| YOLOv4 | CSPDarknet53 | 16 * 8 | 608 | MS COCO 2017 | 46.1 | 337.25 | 27.6M | [yaml](./configs/yolov4/yolov4.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-64b8506f-910v2.ckpt) | +| YOLOv3 | Darknet53 | 16 * 8 | 640 | MS COCO 2017 | 46.6 | 396.60 | 61.9M | [yaml](./configs/yolov3/yolov3.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-81895f09-910v2.ckpt) | +| YOLOX | S | 8 * 8 | 640 | MS COCO 2017 | 41.0 | 242.15 | 9.0M | [yaml](./configs/yolox/yolox-s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-cebd0183-910v2.ckpt) | +
+ +## Segmentation +
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Mask mAP (%) | Params | Recipe | Download | +|------------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv8-seg | X | 16 * 8 | 640 | MS COCO 2017 | 52.5 | 42.9 | 71.8M | [yaml](./configs/yolov8/seg/yolov8x-seg.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x-seg_300e_mAP_mask_429-b4920557.ckpt) | +
+ +## Depoly inference + +- See [support list](./deploy/README.md) + +## Notes +- Box mAP: Accuracy reported on the validation set. diff --git a/configs/yolov3/README.md b/configs/yolov3/README.md index 4ce34a30..03afa3df 100644 --- a/configs/yolov3/README.md +++ b/configs/yolov3/README.md @@ -11,18 +11,26 @@ We present some updates to YOLO! We made a bunch of little design changes to mak ## Results -
+
+performance tested on Ascend 910(8p) with graph mode -| Name | Scale | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|-----------|----------|-----------|--------------|-------------|---------|--------|-----------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------| -| YOLOv3 | Darknet53 | D910x8-G | 640 | MS COCO 2017 | 45.5 | 61.9M | 156.4G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov3/yolov3.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-adfb27af.ckpt) | +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv3 | Darknet53 | 16 * 8 | 640 | MS COCO 2017 | 45.5 | 61.9M | [yaml](./configs/yolov3/yolov3.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-adfb27af.ckpt) | +
+ +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv3 | Darknet53 | 16 * 8 | 640 | MS COCO 2017 | 46.6 | 396.60 | 61.9M | [yaml](./configs/yolov3/yolov3.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov3/yolov3-darknet53_300e_mAP455-81895f09-910v2.ckpt) | +
-

#### Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. - Box mAP: Accuracy reported on the validation set. - We referred to a commonly used third-party [YOLOv3](https://github.com/ultralytics/yolov3) implementation. diff --git a/configs/yolov4/README.md b/configs/yolov4/README.md index e4423cbd..2adb8317 100644 --- a/configs/yolov4/README.md +++ b/configs/yolov4/README.md @@ -25,19 +25,27 @@ AP (65.7% AP50) for the MS COCO dataset at a realtime speed of 65 FPS on Tesla V ## Results -
+
+performance tested on Ascend 910(8p) with graph mode -| Name | Scale | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|--------------|----------|-----------|--------------|-------------|--------|-------|------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------| -| YOLOv4 | CSPDarknet53 | D910x8-G | 608 | MS COCO 2017 | 45.4 | 27.6M | 52G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov4/yolov4.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-50172f93.ckpt) | -| YOLOv4 | CSPDarknet53(silu) | D910x8-G | 608 | MS COCO 2017 | 45.8 | 27.6M | 52G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov4/yolov4-silu.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_silu_320e_map458-bdfc3205.ckpt) | +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv4 | CSPDarknet53 | 16 * 8 | 608 | MS COCO 2017 | 45.4 | 27.6M | [yaml](./configs/yolov4/yolov4.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-50172f93.ckpt) | +| YOLOv4 | CSPDarknet53(silu) | 16 * 8 | 608 | MS COCO 2017 | 45.8 | 27.6M | [yaml](./configs/yolov4/yolov4-silu.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_silu_320e_map458-bdfc3205.ckpt) | +
+ +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv4 | CSPDarknet53 | 16 * 8 | 608 | MS COCO 2017 | 46.1 | 337.25 | 27.6M | [yaml](./configs/yolov4/yolov4.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov4/yolov4-cspdarknet53_320e_map454-64b8506f-910v2.ckpt) | +
-

#### Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. - Box mAP: Accuracy reported on the validation set. ## Quick Start diff --git a/configs/yolov5/README.md b/configs/yolov5/README.md index 1aa2345f..0b751a0c 100644 --- a/configs/yolov5/README.md +++ b/configs/yolov5/README.md @@ -8,22 +8,31 @@ YOLOv5 is a family of object detection architectures and models pretrained on th ## Results -
+
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv5 | N | 32 * 8 | 640 | MS COCO 2017 | 27.3 | 1.9M | [yaml](./configs/yolov5/yolov5n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-9b16bd7b.ckpt) | +| YOLOv5 | S | 32 * 8 | 640 | MS COCO 2017 | 37.6 | 7.2M | [yaml](./configs/yolov5/yolov5s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-860bcf3b.ckpt) | +| YOLOv5 | M | 32 * 8 | 640 | MS COCO 2017 | 44.9 | 21.2M | [yaml](./configs/yolov5/yolov5m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5m_300e_mAP449-e7bbf695.ckpt) | +| YOLOv5 | L | 32 * 8 | 640 | MS COCO 2017 | 48.5 | 46.5M | [yaml](./configs/yolov5/yolov5l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5l_300e_mAP485-a28bce73.ckpt) | +| YOLOv5 | X | 16 * 8 | 640 | MS COCO 2017 | 50.5 | 86.7M | [yaml](./configs/yolov5/yolov5x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5x_300e_mAP505-97d36ddc.ckpt) | +
-| Name | Scale | Arch | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|-------|----------|----------|-----------|--------------|-------------|--------|--------|-----------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------| -| YOLOv5 | N | P5 | D910x8-G | 640 | MS COCO 2017 | 27.3 | 1.9M | 4.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-9b16bd7b.ckpt) | -| YOLOv5 | S | P5 | D910x8-G | 640 | MS COCO 2017 | 37.6 | 7.2M | 16.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-860bcf3b.ckpt) | -| YOLOv5 | M | P5 | D910x8-G | 640 | MS COCO 2017 | 44.9 | 21.2M | 49.0G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5m_300e_mAP449-e7bbf695.ckpt) | -| YOLOv5 | L | P5 | D910x8-G | 640 | MS COCO 2017 | 48.5 | 46.5M | 109.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5l_300e_mAP485-a28bce73.ckpt) | -| YOLOv5 | X | P5 | D910x8-G | 640 | MS COCO 2017 | 50.5 | 86.7M | 205.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov5/yolov5x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov5/yolov5x_300e_mAP505-97d36ddc.ckpt) | +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv5 | N | 32 * 8 | 640 | MS COCO 2017 | 27.4 | 736.08 | 1.9M | [yaml](./configs/yolov5/yolov5n.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov5/yolov5n_300e_mAP273-bedf9a93-910v2.ckpt) | +| YOLOv5 | S | 32 * 8 | 640 | MS COCO 2017 | 37.6 | 787.34 | 7.2M | [yaml](./configs/yolov5/yolov5s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov5/yolov5s_300e_mAP376-df4a45b6-910v2.ckpt) | +
-

#### Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. - Box mAP: Accuracy reported on the validation set. - We refer to the official [YOLOV5](https://github.com/ultralytics/yolov5) to reproduce the P5 series model, and the differences are as follows: 1. We use 8x NPU(Ascend910) for training, and the single-NPU batch size is 32. This is different from the official code. diff --git a/configs/yolov7/README.md b/configs/yolov7/README.md index e288d8d3..f8ff18b7 100644 --- a/configs/yolov7/README.md +++ b/configs/yolov7/README.md @@ -11,15 +11,24 @@ YOLOv7 surpasses all known object detectors in both speed and accuracy in the ra ## Results -
+
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv7 | Tiny | 16 * 8 | 640 | MS COCO 2017 | 37.5 | 6.2M | [yaml](./configs/yolov7/yolov7-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-d8972c94.ckpt) | +| YOLOv7 | L | 16 * 8 | 640 | MS COCO 2017 | 50.8 | 36.9M | [yaml](./configs/yolov7/yolov7.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7_300e_mAP508-734ac919.ckpt) | +| YOLOv7 | X | 12 * 8 | 640 | MS COCO 2017 | 52.4 | 71.3M | [yaml](./configs/yolov7/yolov7-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-x_300e_mAP524-e2f58741.ckpt) | +
-| Name | Scale | Arch | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|-------|------|----------|-----------|--------------|-------------|--------|--------|-----------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------| -| YOLOv7 | Tiny | P5 | D910x8-G | 640 | MS COCO 2017 | 37.5 | 6.2M | 13.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-d8972c94.ckpt) | -| YOLOv7 | L | P5 | D910x8-G | 640 | MS COCO 2017 | 50.8 | 36.9M | 104.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7_300e_mAP508-734ac919.ckpt) | -| YOLOv7 | X | P5 | D910x8-G | 640 | MS COCO 2017 | 52.4 | 71.3M | 189.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov7/yolov7-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov7/yolov7-x_300e_mAP524-e2f58741.ckpt) | +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv7 | Tiny | 16 * 8 | 640 | MS COCO 2017 | 37.5 | 496.21 | 6.2M | [yaml](./configs/yolov7/yolov7-tiny.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov7/yolov7-tiny_300e_mAP375-1d2ddf4b-910v2.ckpt) | +
-

#### Notes diff --git a/configs/yolov8/README.md b/configs/yolov8/README.md index eb227e55..2ae515aa 100644 --- a/configs/yolov8/README.md +++ b/configs/yolov8/README.md @@ -11,31 +11,39 @@ Ultralytics YOLOv8, developed by Ultralytics, is a cutting-edge, state-of-the-ar ### Detection -
+
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOv8 | N | 16 * 8 | 640 | MS COCO 2017 | 37.2 | 3.2M | [yaml](./configs/yolov8/yolov8n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-cc07f5bd.ckpt) | +| YOLOv8 | S | 16 * 8 | 640 | MS COCO 2017 | 44.6 | 11.2M | [yaml](./configs/yolov8/yolov8s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-3086f0c9.ckpt) | +| YOLOv8 | M | 16 * 8 | 640 | MS COCO 2017 | 50.5 | 25.9M | [yaml](./configs/yolov8/yolov8m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-m_500e_mAP505-8ff7a728.ckpt) | +| YOLOv8 | L | 16 * 8 | 640 | MS COCO 2017 | 52.8 | 43.7M | [yaml](./configs/yolov8/yolov8l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-l_500e_mAP528-6e96d6bb.ckpt) | +| YOLOv8 | X | 16 * 8 | 640 | MS COCO 2017 | 53.7 | 68.2M | [yaml](./configs/yolov8/yolov8x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x_500e_mAP537-b958e1c7.ckpt) | +
-| Name | Scale | Arch | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|-------|------|----------|-----------|--------------|-------------|--------|--------|-------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------| -| YOLOv8 | N | P5 | D910x8-G | 640 | MS COCO 2017 | 37.2 | 3.2M | 8.7G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8n.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-cc07f5bd.ckpt) | -| YOLOv8 | S | P5 | D910x8-G | 640 | MS COCO 2017 | 44.6 | 11.2M | 28.6G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-3086f0c9.ckpt) | -| YOLOv8 | M | P5 | D910x8-G | 640 | MS COCO 2017 | 50.5 | 25.9M | 78.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-m_500e_mAP505-8ff7a728.ckpt) | -| YOLOv8 | L | P5 | D910x8-G | 640 | MS COCO 2017 | 52.8 | 43.7M | 165.2G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-l_500e_mAP528-6e96d6bb.ckpt) | -| YOLOv8 | X | P5 | D910x8-G | 640 | MS COCO 2017 | 53.7 | 68.2M | 257.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/yolov8x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x_500e_mAP537-b958e1c7.ckpt) | +
+performance tested on Ascend 910*(8p) -
+| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv8 | N | 16 * 8 | 640 | MS COCO 2017 | 37.3 | 373.55 | 3.2M | [yaml](./configs/yolov8/yolov8n.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov8/yolov8-n_500e_mAP372-0e737186-910v2.ckpt) | +| YOLOv8 | S | 16 * 8 | 640 | MS COCO 2017 | 44.7 | 365.53 | 11.2M | [yaml](./configs/yolov8/yolov8s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolov8/yolov8-s_500e_mAP446-fae4983f-910v2.ckpt) | + ### Segmentation -
+
+performance tested on Ascend 910(8p) with graph mode -| Name | Scale | Arch | Context | ImageSize | Dataset | Box mAP (%) | Mask mAP (%) | Params | FLOPs | Recipe | Download | -|------------|-------|------|----------|-----------|--------------|-------------|--------------|--------|--------|---------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------| -| YOLOv8-seg | X | P5 | D910x8-G | 640 | MS COCO 2017 | 52.5 | 42.9 | 71.8M | 344.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolov8/seg/yolov8x-seg.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x-seg_300e_mAP_mask_429-b4920557.ckpt) | - -
+| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Mask mAP (%) | Params | Recipe | Download | +|------------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOv8-seg | X | 16 * 8 | 640 | MS COCO 2017 | 52.5 | 42.9 | 71.8M | [yaml](./configs/yolov8/seg/yolov8x-seg.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolov8/yolov8-x-seg_300e_mAP_mask_429-b4920557.ckpt) | + ### Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. - Box mAP: Accuracy reported on the validation set. - We refer to the official [YOLOV8](https://github.com/ultralytics/ultralytics) to reproduce the P5 series model. diff --git a/configs/yolox/README.md b/configs/yolox/README.md index 2b27c0c5..bc01181b 100644 --- a/configs/yolox/README.md +++ b/configs/yolox/README.md @@ -8,24 +8,32 @@ YOLOX is a new high-performance detector with some experienced improvements to Y ## Results -
- -| Name | Scale | Context | ImageSize | Dataset | Box mAP (%) | Params | FLOPs | Recipe | Download | -|--------|-----------|----------|-----------|--------------|-------------|--------|--------|----------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------| -| YOLOX | N | D910x8-G | 416 | MS COCO 2017 | 24.1 | 0.9M | 1.1G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-nano.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-n_300e_map241-ec9815e3.ckpt) | -| YOLOX | Tiny | D910x8-G | 416 | MS COCO 2017 | 33.3 | 5.1M | 6.5G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-tiny_300e_map333-e5ae3a2e.ckpt) | -| YOLOX | S | D910x8-G | 640 | MS COCO 2017 | 40.7 | 9.0M | 26.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-0983e07f.ckpt) | -| YOLOX | M | D910x8-G | 640 | MS COCO 2017 | 46.7 | 25.3M | 73.8G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-m_300e_map467-1db321ee.ckpt) | -| YOLOX | L | D910x8-G | 640 | MS COCO 2017 | 49.2 | 54.2M | 155.6G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-l_300e_map492-52a4ab80.ckpt) | -| YOLOX | X | D910x8-G | 640 | MS COCO 2017 | 51.6 | 99.1M | 281.9G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-x_300e_map516-52216d90.ckpt) | -| YOLOX | Darknet53 | D910x8-G | 640 | MS COCO 2017 | 47.7 | 63.7M | 185.3G | [yaml](https://github.com/mindspore-lab/mindyolo/blob/master/configs/yolox/yolox-darknet53.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-darknet53_300e_map477-b5fcaba9.ckpt) | +
+performance tested on Ascend 910(8p) with graph mode + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | +| YOLOX | N | 8 * 8 | 416 | MS COCO 2017 | 24.1 | 0.9M | [yaml](./configs/yolox/yolox-nano.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-n_300e_map241-ec9815e3.ckpt) | +| YOLOX | Tiny | 8 * 8 | 416 | MS COCO 2017 | 33.3 | 5.1M | [yaml](./configs/yolox/yolox-tiny.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-tiny_300e_map333-e5ae3a2e.ckpt) | +| YOLOX | S | 8 * 8 | 640 | MS COCO 2017 | 40.7 | 9.0M | [yaml](./configs/yolox/yolox-s.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-0983e07f.ckpt) | +| YOLOX | M | 8 * 8 | 640 | MS COCO 2017 | 46.7 | 25.3M | [yaml](./configs/yolox/yolox-m.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-m_300e_map467-1db321ee.ckpt) | +| YOLOX | L | 8 * 8 | 640 | MS COCO 2017 | 49.2 | 54.2M | [yaml](./configs/yolox/yolox-l.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-l_300e_map492-52a4ab80.ckpt) | +| YOLOX | X | 8 * 8 | 640 | MS COCO 2017 | 51.6 | 99.1M | [yaml](./configs/yolox/yolox-x.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-x_300e_map516-52216d90.ckpt) | +| YOLOX | Darknet53 | 8 * 8 | 640 | MS COCO 2017 | 47.7 | 63.7M | [yaml](./configs/yolox/yolox-darknet53.yaml) | [weights](https://download.mindspore.cn/toolkits/mindyolo/yolox/yolox-darknet53_300e_map477-b5fcaba9.ckpt) | +
+ +
+performance tested on Ascend 910*(8p) + +| Name | Scale | BatchSize | ImageSize | Dataset | Box mAP (%) | ms/step | Params | Recipe | Download | +|--------| :---: | :---: | :---: |--------------| :---: | :---: | :---: | :---: | :---: | +| YOLOX | S | 8 * 8 | 640 | MS COCO 2017 | 41.0 | 242.15 | 9.0M | [yaml](./configs/yolox/yolox-s.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindyolo/yolox/yolox-s_300e_map407-cebd0183-910v2.ckpt) | +
-

#### Notes -- Context: Training context denoted as {device}x{pieces}-{MS mode}, where mindspore mode can be G - graph mode or F - pynative mode with ms function. For example, D910x8-G is for training on 8 pieces of Ascend 910 NPU using graph mode. - Box mAP: Accuracy reported on the validation set. - We refer to the official [YOLOX](https://github.com/Megvii-BaseDetection/YOLOX) to reproduce the results. diff --git a/docs/en/modelzoo.md b/docs/en/modelzoo.md index 85365b4c..9e549c51 100644 --- a/docs/en/modelzoo.md +++ b/docs/en/modelzoo.md @@ -6,4 +6,4 @@ hide: # Model Zoo -{% include-markdown "../../MODEL_ZOO.md" %} +{% include-markdown "../../benchmark_results.md" %} diff --git a/docs/zh/modelzoo.md b/docs/zh/modelzoo.md index 9f2f81a9..21936b0b 100644 --- a/docs/zh/modelzoo.md +++ b/docs/zh/modelzoo.md @@ -6,4 +6,4 @@ hide: # 模型仓库 -{% include-markdown "../../MODEL_ZOO.md" %} +{% include-markdown "../../benchmark_results.md" %} diff --git a/examples/finetune_SHWD/README.md b/examples/finetune_SHWD/README.md index 1f31bbd6..89ecdec5 100644 --- a/examples/finetune_SHWD/README.md +++ b/examples/finetune_SHWD/README.md @@ -100,7 +100,7 @@ optimizer: 参数继承关系和参数说明可参考[configuration_CN.md](../../tutorials/configuration_CN.md)。 #### 下载预训练模型 -可选用MindYOLO提供的[MODEL_ZOO](../../MODEL_ZOO.md)作为自定义数据集的预训练模型，预训练模型在COCO数据集上已经有较好的精度表现，相比从头训练，加载预训练模型一般会拥有更快的收敛速度以及更高的最终精度，并且大概率能避免初始化不当导致的梯度消失、梯度爆炸等问题。 +可选用MindYOLO提供的[模型仓库](../../benchmark_results.md)作为自定义数据集的预训练模型，预训练模型在COCO数据集上已经有较好的精度表现，相比从头训练，加载预训练模型一般会拥有更快的收敛速度以及更高的最终精度，并且大概率能避免初始化不当导致的梯度消失、梯度爆炸等问题。 自定义数据集类别数通常与COCO数据集不一致，MindYOLO中各模型的检测头head结构跟数据集类别数有关，直接将预训练模型导入可能会因为shape不一致而导入失败，可以在yaml配置文件中设置strict_load参数为False，MindYOLO将自动舍弃shape不一致的参数，并抛出该module参数并未导入的告警 #### 模型微调(Finetune) diff --git a/examples/finetune_car_detection/README.md b/examples/finetune_car_detection/README.md index fa1bf819..4370f02e 100644 --- a/examples/finetune_car_detection/README.md +++ b/examples/finetune_car_detection/README.md @@ -57,7 +57,7 @@ bdd_ud ## 实验过程 选择模型时，尝试着先用较小的模型yolov7-tiny在普通的UA-DETRAC数据集上来完成实验并测试结果精度，之后再尝试更复杂的模型如yolov7l在整合后的数据集上来进行实验，识别更多的类别达到更好的效果。 -预训练模型权重参数可以在[MODEL_ZOO.md](../../MODEL_ZOO.md)中下载得到。 +预训练模型权重参数可以在[模型仓库](../../benchmark_results.md)中下载得到。 ## yolov7-tiny实验过程 ### 编写yaml配置文件