Bump version to v0.1.3

open-mmlab · Nov 10, 2022 · 178b0bf · 178b0bf
2 parents 0b48313 + abcde76
commit 178b0bf
Show file tree

Hide file tree

Showing 47 changed files with 753 additions and 323 deletions.
diff --git a/.circleci/test.yml b/.circleci/test.yml
@@ -66,7 +66,7 @@ jobs:
           name: Install mmyolo dependencies
           command: |
             pip install -U openmim
-            mim install 'mmengine >= 0.3.0'
+            mim install 'mmengine >= 0.3.1'
             mim install 'mmcv >= 2.0.0rc1'
             pip install git+https://github.com/open-mmlab/[email protected]
             pip install -r requirements/albu.txt
@@ -124,7 +124,7 @@ jobs:
           name: Install mmyolo dependencies
           command: |
             docker exec mmyolo pip install -U openmim
-            docker exec mmyolo mim install 'mmengine >= 0.3.0'
+            docker exec mmyolo mim install 'mmengine >= 0.3.1'
             docker exec mmyolo mim install 'mmcv >= 2.0.0rc1'
             docker exec mmyolo pip install -e /mmdetection
             docker exec mmyolo pip install -r requirements/albu.txt

diff --git a/README.md b/README.md
@@ -62,13 +62,20 @@ The master branch works with **PyTorch 1.6+**.
 
   MMYOLO decomposes the framework into different components where users can easily customize a model by combining different modules with various training and testing strategies.
 
-<img src="https://user-images.githubusercontent.com/27466624/190986949-01414a91-baae-4228-8828-c59db58dcf36.jpg" alt="BaseModule"/>
-  The figure is contributed by RangeKing@GitHub, thank you very much!
+<img src="https://user-images.githubusercontent.com/27466624/199999337-0544a4cb-3cbd-4f3e-be26-bcd9e74db7ff.jpg" alt="BaseModule-P5"/>
+  The figure above is contributed by RangeKing@GitHub, thank you very much!
+
+And the figure of P6 model is in [model_design.md](docs/en/algorithm_descriptions/model_design.md).
 
 </details>
 
 ## What's New
 
+💎 **v0.1.3** was released on 10/11/2022:
+
+1. Fix training failure when saving best weights based on mmengine 0.3.1
+2. Fix `add_dump_metric` error based on mmdet 3.0.0rc3
+
 💎 **v0.1.2** was released on 3/11/2022:
 
 1. Support [YOLOv5/YOLOv6/YOLOX/RTMDet deployments](https://github.com/open-mmlab/mmyolo/blob/main/configs/deploy) for ONNXRuntime and TensorRT
@@ -88,9 +95,9 @@ MMYOLO relies on PyTorch, MMCV, MMEngine, and MMDetection. Below are quick steps
 conda create -n open-mmlab python=3.8 pytorch==1.10.1 torchvision==0.11.2 cudatoolkit=11.3 -c pytorch -y
 conda activate open-mmlab
 pip install openmim
-mim install "mmengine>=0.3.0"
+mim install "mmengine>=0.3.1"
 mim install "mmcv>=2.0.0rc1,<2.1.0"
-mim install "mmdet>=3.0.0rc2,<3.1.0"
+mim install "mmdet>=3.0.0rc3,<3.1.0"
 git clone https://github.com/open-mmlab/mmyolo.git
 cd mmyolo
 # Install albumentations
@@ -135,6 +142,7 @@ For different parts from MMDetection, we have also prepared user guides and adva
 
   - [Data flow](docs/en/advanced_guides/data_flow.md)
   - [How to](docs/en/advanced_guides/how_to.md)
+  - [Plugins](docs/en/advanced_guides/plugins.md)
 
 ## Overview of Benchmark and Model Zoo
 

diff --git a/README_zh-CN.md b/README_zh-CN.md
@@ -62,13 +62,20 @@ MMYOLO 是一个基于 PyTorch 和 MMDetection 的 YOLO 系列算法开源工具
 
   MMYOLO 将框架解耦成不同的模块组件，通过组合不同的模块和训练测试策略，用户可以便捷地构建自定义模型。
 
-<img src="https://user-images.githubusercontent.com/27466624/190986949-01414a91-baae-4228-8828-c59db58dcf36.jpg" alt="基类"/>
+<img src="https://user-images.githubusercontent.com/27466624/199999337-0544a4cb-3cbd-4f3e-be26-bcd9e74db7ff.jpg" alt="基类-P5"/>
   图为 RangeKing@GitHub 提供，非常感谢！
 
+P6 模型图详见 [model_design.md](docs/zh_CN/algorithm_descriptions/model_design.md)。
+
 </details>
 
 ## 最新进展
 
+💎 **v0.1.3** 版本已经在 2022.11.10 发布：
+
+1. 基于 mmengine 0.3.1 修复保存最好权重时训练失败问题
+2. 基于 mmdet 3.0.0rc3 修复 `add_dump_metric` 报错 (#253)
+
 💎 **v0.1.2** 版本已经在 2022.11.3 发布：
 
 1. 支持 ONNXRuntime 和 TensorRT 的 [YOLOv5/YOLOv6/YOLOX/RTMDet 部署](https://github.com/open-mmlab/mmyolo/blob/main/configs/deploy)
@@ -97,9 +104,9 @@ MMYOLO 依赖 PyTorch, MMCV, MMEngine 和 MMDetection，以下是安装的简要
 conda create -n open-mmlab python=3.8 pytorch==1.10.1 torchvision==0.11.2 cudatoolkit=11.3 -c pytorch -y
 conda activate open-mmlab
 pip install openmim
-mim install "mmengine>=0.3.0"
+mim install "mmengine>=0.3.1"
 mim install "mmcv>=2.0.0rc1,<2.1.0"
-mim install "mmdet>=3.0.0rc2,<3.1.0"
+mim install "mmdet>=3.0.0rc3,<3.1.0"
 git clone https://github.com/open-mmlab/mmyolo.git
 cd mmyolo
 # Install albumentations
@@ -145,6 +152,7 @@ MMYOLO 用法和 MMDetection 几乎一致，所有教程都是通用的，你也
 
   - [数据流](docs/zh_cn/advanced_guides/data_flow.md)
   - [How to](docs/zh_cn/advanced_guides/how_to.md)
+  - [插件](docs/zh_cn/advanced_guides/plugins.md)
 
 - [解读文章和资源汇总](docs/zh_cn/article.md)
 

diff --git a/configs/rtmdet/README.md b/configs/rtmdet/README.md
@@ -12,13 +12,13 @@ Our tech-report will be released soon.
 
 ## Results and Models
 
-|  Backbone   | size | SyncBN | ox AP | Params(M) | FLOPS(G) | TRT-FP16-Latency(ms) |                      Config                       |                                                                                                                                                   Download                                                                                                                                                    |
-| :---------: | :--: | :----: | ----: | :-------: | :------: | :------------------: | :-----------------------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
-| RTMDet-tiny | 640  |  Yes   |  40.9 |    4.8    |   8.1    |         0.98         | [config](./rtmdet_tiny_syncbn_8xb32-300e_coco.py) | [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_tiny_syncbn_8xb32-300e_coco/rtmdet_tiny_syncbn_8xb32-300e_coco_20220902_112414-259f3241.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_tiny_8xb32-300e_coco/rtmdet_tiny_8xb32-300e_coco_20220902_112414.log.json) |
-|  RTMDet-s   | 640  |  Yes   |  44.5 |   8.89    |   14.8   |         1.22         |  [config](./rtmdet_s_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_s_syncbn_8xb32-300e_coco/rtmdet_s_syncbn_8xb32-300e_coco_20220905_161602-fd1cacb9.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_s_8xb32-300e_coco/rtmdet_s_8xb32-300e_coco_20220905_161602.log.json)       |
-|  RTMDet-m   | 640  |  Yes   |  49.1 |   24.71   |  39.27   |         1.62         |  [config](./rtmdet_m_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_m_syncbn_8xb32-300e_coco/rtmdet_m_syncbn_8xb32-300e_coco_20220924_132959-d9f2e90d.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_m_8xb32-300e_coco/rtmdet_m_8xb32-300e_coco_20220924_132959.log.json)       |
-|  RTMDet-l   | 640  |  Yes   |  51.3 |   52.3    |  80.23   |         2.44         |  [config](./rtmdet_l_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_l_syncbn_8xb32-300e_coco/rtmdet_l_syncbn_8xb32-300e_coco_20220926_150401-40c754b5.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_l_8xb32-300e_coco/rtmdet_l_8xb32-300e_coco_20220926_150401.log.json)       |
-|  RTMDet-x   | 640  |  Yes   |  52.6 |   94.86   |  141.67  |         3.10         |  [config](./rtmdet_x_syncbn_8xb32-300e_coco.py)   |                                                                                                                                           [model](<>) \| [log](<>)                                                                                                                                            |
+|  Backbone   | size | SyncBN | box AP | Params(M) | FLOPS(G) | TRT-FP16-Latency(ms) |                      Config                       |                                                                                                                                                   Download                                                                                                                                                    |
+| :---------: | :--: | :----: | -----: | :-------: | :------: | :------------------: | :-----------------------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
+| RTMDet-tiny | 640  |  Yes   |   40.9 |    4.8    |   8.1    |         0.98         | [config](./rtmdet_tiny_syncbn_8xb32-300e_coco.py) | [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_tiny_syncbn_8xb32-300e_coco/rtmdet_tiny_syncbn_8xb32-300e_coco_20220902_112414-259f3241.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_tiny_8xb32-300e_coco/rtmdet_tiny_8xb32-300e_coco_20220902_112414.log.json) |
+|  RTMDet-s   | 640  |  Yes   |   44.5 |   8.89    |   14.8   |         1.22         |  [config](./rtmdet_s_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_s_syncbn_8xb32-300e_coco/rtmdet_s_syncbn_8xb32-300e_coco_20220905_161602-fd1cacb9.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_s_8xb32-300e_coco/rtmdet_s_8xb32-300e_coco_20220905_161602.log.json)       |
+|  RTMDet-m   | 640  |  Yes   |   49.1 |   24.71   |  39.27   |         1.62         |  [config](./rtmdet_m_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_m_syncbn_8xb32-300e_coco/rtmdet_m_syncbn_8xb32-300e_coco_20220924_132959-d9f2e90d.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_m_8xb32-300e_coco/rtmdet_m_8xb32-300e_coco_20220924_132959.log.json)       |
+|  RTMDet-l   | 640  |  Yes   |   51.3 |   52.3    |  80.23   |         2.44         |  [config](./rtmdet_l_syncbn_8xb32-300e_coco.py)   |       [model](https://download.openmmlab.com/mmyolo/v0/rtmdet/rtmdet_l_syncbn_8xb32-300e_coco/rtmdet_l_syncbn_8xb32-300e_coco_20220926_150401-40c754b5.pth) \| [log](https://download.openmmlab.com/mmdetection/v3.0/rtmdet/rtmdet_l_8xb32-300e_coco/rtmdet_l_8xb32-300e_coco_20220926_150401.log.json)       |
+|  RTMDet-x   | 640  |  Yes   |   52.6 |   94.86   |  141.67  |         3.10         |  [config](./rtmdet_x_syncbn_8xb32-300e_coco.py)   |                                                                                                                                           [model](<>) \| [log](<>)                                                                                                                                            |
 
 **Note**:
 

diff --git a/demo/featmap_vis_demo.py b/demo/featmap_vis_demo.py
@@ -1,39 +1,45 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import argparse
+import os
 from typing import Sequence
 
 import mmcv
-import numpy as np
 from mmdet.apis import inference_detector, init_detector
 from mmengine import Config, DictAction
+from mmengine.utils import ProgressBar
 
 from mmyolo.registry import VISUALIZERS
 from mmyolo.utils import register_all_modules
+from mmyolo.utils.misc import auto_arrange_images, get_file_list
 
 
 # TODO: Refine
 def parse_args():
-    parser = argparse.ArgumentParser(description='Visualize Featmaps')
-    parser.add_argument('img', help='Image file')
+    parser = argparse.ArgumentParser(description='Visualize feature map')
+    parser.add_argument(
+        'img', help='Image path, include image file, dir and URL.')
     parser.add_argument('config', help='Config file')
     parser.add_argument('checkpoint', help='Checkpoint file')
+    parser.add_argument(
+        '--out-dir', default='./output', help='Path to output file')
     parser.add_argument(
         '--target-layers',
         default=['backbone'],
         nargs='+',
         type=str,
-        help='The target layers to get featmap, if not set, the tool will '
+        help='The target layers to get feature map, if not set, the tool will '
         'specify the backbone')
     parser.add_argument(
         '--preview-model',
         default=False,
         action='store_true',
         help='To preview all the model layers')
-    parser.add_argument('--out-file', default=None, help='Path to output file')
     parser.add_argument(
         '--device', default='cuda:0', help='Device used for inference')
     parser.add_argument(
         '--score-thr', type=float, default=0.3, help='Bbox score threshold')
+    parser.add_argument(
+        '--show', action='store_true', help='Show the featmap results')
     parser.add_argument(
         '--channel-reduction',
         default='select_max',
@@ -88,28 +94,9 @@ def release(self):
             handle.remove()
 
 
-def auto_arrange_imgs(imgs):
-    len_img = len(imgs)
-    col = 2
-    if len_img <= col:
-        imgs = np.concatenate(imgs, axis=1)
-    else:
-        row = len_img // col + 1
-        fill_img_list = [np.ones(imgs[0].shape, dtype=np.uint8) * 255] * (
-            row * col - len_img)
-        imgs.extend(fill_img_list)
-        merge_imgs_col = []
-        for i in range(row):
-            start = col * i
-            end = col * (i + 1)
-            merge_col = np.hstack(imgs[start:end])
-            merge_imgs_col.append(merge_col)
-
-        imgs = np.vstack(merge_imgs_col)
-    return imgs
-
-
-def main(args):
+def main():
+    args = parse_args()
+
     # register all modules in mmdet into the registries
     register_all_modules()
 
@@ -124,9 +111,13 @@ def main(args):
 
     model = init_detector(args.config, args.checkpoint, device=args.device)
 
+    if not os.path.exists(args.out_dir) and not args.show:
+        os.mkdir(args.out_dir)
+
     if args.preview_model:
         print(model)
-        print('\n Please remove `--preview-model` to get the AM.')
+        print('\n This flag is only show model, if you want to continue, '
+              'please remove `--preview-model` to get the feature map.')
         return
 
     target_layers = []
@@ -143,11 +134,11 @@ def main(args):
     visualizer = VISUALIZERS.build(model.cfg.visualizer)
     visualizer.dataset_meta = model.dataset_meta
 
-    images = args.img
-    if not isinstance(images, list):
-        images = [images]
+    # get file list
+    image_list, source_type = get_file_list(args.img)
 
-    for image_path in images:
+    progress_bar = ProgressBar(len(image_list))
+    for image_path in image_list:
         result, featmaps = activations_wrapper(image_path)
         if not isinstance(featmaps, Sequence):
             featmaps = [featmaps]
@@ -159,10 +150,16 @@ def main(args):
             else:
                 flatten_featmaps.append(featmap)
 
-        # show the results
-        img = mmcv.imread(args.img)
+        img = mmcv.imread(image_path)
         img = mmcv.imconvert(img, 'bgr', 'rgb')
 
+        if source_type['is_dir']:
+            filename = os.path.relpath(image_path, args.img).replace('/', '_')
+        else:
+            filename = os.path.basename(image_path)
+        out_file = None if args.show else os.path.join(args.out_dir, filename)
+
+        # show the results
         shown_imgs = []
         visualizer.add_datasample(
             'result',
@@ -184,16 +181,20 @@ def main(args):
                 arrangement=args.arrangement)
             shown_imgs.append(shown_img)
 
-        shown_imgs = auto_arrange_imgs(shown_imgs)
+        shown_imgs = auto_arrange_images(shown_imgs)
 
-        if args.out_file is not None:
-            mmcv.imwrite(shown_imgs[..., ::-1], args.out_file)
-        else:
+        progress_bar.update()
+        if out_file:
+            mmcv.imwrite(shown_imgs[..., ::-1], out_file)
+
+        if args.show:
             visualizer.show(shown_imgs)
 
+    print(f'All done!'
+          f'\nResults have been saved at {os.path.abspath(args.out_dir)}')
+
 
 # Please refer to the usage tutorial:
 # https://github.com/open-mmlab/mmyolo/blob/main/docs/zh_cn/user_guides/visualization.md # noqa
 if __name__ == '__main__':
-    args = parse_args()
-    main(args)
+    main()
diff --git a/demo/image_demo.py b/demo/image_demo.py
@@ -1,20 +1,15 @@
 # Copyright (c) OpenMMLab. All rights reserved.
-import logging
 import os
-import urllib
 from argparse import ArgumentParser
 
 import mmcv
-import torch
 from mmdet.apis import inference_detector, init_detector
 from mmengine.logging import print_log
-from mmengine.utils import ProgressBar, scandir
+from mmengine.utils import ProgressBar
 
 from mmyolo.registry import VISUALIZERS
 from mmyolo.utils import register_all_modules, switch_to_deploy
-
-IMG_EXTENSIONS = ('.jpg', '.jpeg', '.png', '.ppm', '.bmp', '.pgm', '.tif',
-                  '.tiff', '.webp')
+from mmyolo.utils.misc import get_file_list
 
 
 def parse_args():
@@ -39,7 +34,9 @@ def parse_args():
     return args
 
 
-def main(args):
+def main():
+    args = parse_args()
+
     # register all modules in mmdet into the registries
     register_all_modules()
 
@@ -49,45 +46,32 @@ def main(args):
     if args.deploy:
         switch_to_deploy(model)
 
+    if not os.path.exists(args.out_dir) and not args.show:
+        os.mkdir(args.out_dir)
+
     # init visualizer
     visualizer = VISUALIZERS.build(model.cfg.visualizer)
     visualizer.dataset_meta = model.dataset_meta
 
-    is_dir = os.path.isdir(args.img)
-    is_url = args.img.startswith(('http:/', 'https:/'))
-    is_file = os.path.splitext(args.img)[-1] in (IMG_EXTENSIONS)
-
-    files = []
-    if is_dir:
-        # when input source is dir
-        for file in scandir(args.img, IMG_EXTENSIONS, recursive=True):
-            files.append(os.path.join(args.img, file))
-    elif is_url:
-        # when input source is url
-        filename = os.path.basename(
-            urllib.parse.unquote(args.img).split('?')[0])
-        torch.hub.download_url_to_file(args.img, filename)
-        files = [os.path.join(os.getcwd(), filename)]
-    elif is_file:
-        # when input source is single image
-        files = [args.img]
-    else:
-        print_log(
-            'Cannot find image file.', logger='current', level=logging.WARNING)
+    # get file list
+    files, source_type = get_file_list(args.img)
 
     # start detector inference
     progress_bar = ProgressBar(len(files))
     for file in files:
         result = inference_detector(model, file)
+
         img = mmcv.imread(file)
         img = mmcv.imconvert(img, 'bgr', 'rgb')
-        if is_dir:
+
+        if source_type['is_dir']:
             filename = os.path.relpath(file, args.img).replace('/', '_')
         else:
             filename = os.path.basename(file)
         out_file = None if args.show else os.path.join(args.out_dir, filename)
+
         visualizer.add_datasample(
-            filename,
+            os.path.basename(out_file),
             img,
             data_sample=result,
             draw_gt=False,
@@ -96,11 +80,11 @@ def main(args):
             out_file=out_file,
             pred_score_thr=args.score_thr)
         progress_bar.update()
+
     if not args.show:
         print_log(
             f'\nResults have been saved at {os.path.abspath(args.out_dir)}')
 
 
 if __name__ == '__main__':
-    args = parse_args()
-    main(args)
+    main()
diff --git a/docker/Dockerfile b/docker/Dockerfile
@@ -26,7 +26,7 @@ RUN apt-get update \
 
 # Install MMEngine , MMCV and MMDet
 RUN pip install --no-cache-dir openmim && \
-    mim install --no-cache-dir "mmengine>=0.3.0" "mmcv>=2.0.0rc1,<2.1.0" "mmdet>=3.0.0rc2,<3.1.0"
+    mim install --no-cache-dir "mmengine>=0.3.1" "mmcv>=2.0.0rc1,<2.1.0" "mmdet>=3.0.0rc3,<3.1.0"
 
 # Install MMYOLO
 RUN git clone https://github.com/open-mmlab/mmyolo.git /mmyolo && \