custon applications

PNightOwlY · Oct 9, 2019 · 84dfbdf · 84dfbdf
1 parent dbb00d8
commit 84dfbdf
Show file tree

Hide file tree

Showing 10 changed files with 65 additions and 73 deletions.
diff --git a/README.md b/README.md
@@ -11,6 +11,8 @@ MMSkeleton is developed on our research project [ST-GCN](https://github.com/yysi
 </p>
 
 ## Updates
+- [2019-10-09] MMSkeleton v0.6 is released.
+- [2019-10-08] Support model zoo.
 - [2019-10-02] Support custom dataset.
 - [2019-09-23] Add video-based pose estimation demo.
 - [2019-08-29] MMSkeleton v0.5 is released.
@@ -25,16 +27,17 @@ MMSkeleton is developed on our research project [ST-GCN](https://github.com/yysi
 - **Multiple tasks**
 
     MMSkeleton addresses to multiple tasks in human understanding, including but not limited to:
-    - [x] build custom skeleton-based dataset: [[CUSTOM_DATASET]](./doc/CUSTOM_DATASET.md)
-    - [x] skeleton-based action recognition: [[ST-GCN]](./doc/START_RECOGNITION.md)
-    - [x] 2D pose estimation: [[START_POSE_ESTIMATION.md]](./doc/START_POSE_ESTIMATION.md)
+    - [x] [skeleton-based action recognition (ST-GCN)](./doc/START_RECOGNITION.md)
+    - [x] [2D pose estimation](./doc/START_POSE_ESTIMATION.md)
     - [ ] skeleton-based action generation
     - [ ] 3D pose estimation
     - [ ] pose tracking
+    - [x] [build custom skeleton-based dataset](./doc/CUSTOM_DATASET.md)
+    - [x] [create your own applications](./doc/CREATE_APPLICATION.md)
 
 ## Getting Started
 
-Please see [GETTING_STARTED.md](./doc/GETTING_STARTED.md) and [START_RECOGNITION.md](./doc/START_RECOGNITION.md) for more details of MMSkeleton.
+Please see [GETTING_STARTED.md](./doc/GETTING_STARTED.md) for more details of MMSkeleton.
 
 ## License
 The project is release under the [Apache 2.0 license](./LICENSE).

diff --git a/configs/recognition/st_gcn/kinetics-skeleton-from-openpose/test.yaml b/configs/recognition/st_gcn/kinetics-skeleton-from-openpose/test.yaml
@@ -29,7 +29,7 @@ processor_cfg:
     type: "datasets.DataPipeline"
     data_source:
       type: "datasets.SkeletonLoader"
-      data_dir: ./data/Kinetics/skeleton-from-openpose/val
+      data_dir: ./data/Kinetics/skeleton-from-openpose/val2014
       num_track: 2
     pipeline:
       - type: "datasets.skeleton.normalize_by_resolution"

diff --git a/configs/recognition/st_gcn/kinetics-skeleton-from-openpose/train.yaml b/configs/recognition/st_gcn/kinetics-skeleton-from-openpose/train.yaml
@@ -34,7 +34,7 @@ processor_cfg:
     - type: "datasets.DataPipeline"
       data_source:
         type: "datasets.SkeletonLoader"
-        data_dir: ./data/Kinetics/skeleton-from-openpose/train
+        data_dir: ./data/Kinetics/skeleton-from-openpose/train2014
         num_track: 2
       pipeline:
         - type: "datasets.skeleton.normalize_by_resolution"
@@ -49,7 +49,7 @@ processor_cfg:
     - type: "datasets.DataPipeline"
       data_source:
         type: "datasets.SkeletonLoader"
-        data_dir: ./data/Kinetics/skeleton-from-openpose/val
+        data_dir: ./data/Kinetics/skeleton-from-openpose/val2014
         num_track: 2
       pipeline:
         - type: "datasets.skeleton.normalize_by_resolution"

diff --git a/doc/CREATE_APPLICATION.md b/doc/CREATE_APPLICATION.md
@@ -0,0 +1,41 @@
+## Create an MMSkeleton Application
+
+MMSkeleton provides various models, datasets, apis, operators for various applications,
+such as pose estimation, human detection, action recognition and dataset building.
+The workflow of a application is defined by a **processor**, which is usually a python function.
+
+In MMSkeleton, an application is corresponded to a configuration file.
+It is a `.json`, `.yaml` or `.py` file including `processor_cfg` field. 
+There is an example:
+
+```yaml
+# yaml
+
+processor_cfg:
+  type: <path to processor module>
+  dataset:
+    type: <path to dataset module>
+    data_path: ./data
+  ...
+
+argparse_cfg:
+  data:
+    bind_to: processor_cfg.dataset.data_path
+    help: the path of data
+  ...
+```
+
+The `processor_cfg` specifies a processor module and its dataset module
+In adittion, the `data_path` argument of the dataset is "./data".
+The `argparse_cfg` create a option argument `data` which is bound to `data_path`.
+
+With this configuration, the application can be started by:
+```shell
+mmskl $CONFIG_FILE --data $DATA_PATH
+```
+
+
+
+
+
+
diff --git a/doc/EXTENSIONS.md b/doc/EXTENSIONS.md
diff --git a/doc/GETTING_STARTED.md b/doc/GETTING_STARTED.md
@@ -44,4 +44,10 @@ mmskl $CONFIG_FILE -h
 ### Example:
 
 See [START_RECOGNITION.md](../doc/START_RECOGNITION.md) for learning how to train a model for skeleton-based action recognitoin.
+
 See [CUSTOM_DATASET](./doc/CUSTOM_DATASET.md) for building your own skeleton-based dataset.
+
+See [CREATE_APPLICATION](./doc/CREATE_APPLICATION.md) for creating your own mmskeleton application.
+
+
+
diff --git a/doc/MMSKL_EXTENSIONS.md b/doc/MMSKL_EXTENSIONS.md
diff --git a/mmskeleton/__init__.py b/mmskeleton/__init__.py
@@ -1 +1,2 @@
+from . import datasets, processor, models, ops, apis
 from .datasets.skeleton import skeleton_process
diff --git a/mmskeleton/deprecated/processor/pseudo.py b/mmskeleton/deprecated/processor/pseudo.py
@@ -5,4 +5,9 @@ def train(model_cfg, dataset_cfg, optimizer):
     model = call_obj(**model_cfg)
     dataset = call_obj(**dataset_cfg)
     print('train a pseudo model...')
-    print('done.')
+    print('done.')
+
+
+def hello_world(times=10):
+    for i in range(times):
+        print('Hello World!')
diff --git a/mmskeleton/processor/__init__.py b/mmskeleton/processor/__init__.py
@@ -0,0 +1 @@
+from . import recognition, pose_demo
Original file line number	Diff line number	Diff line change
		@@ -1 +1,2 @@
		from . import datasets, processor, models, ops, apis
		from .datasets.skeleton import skeleton_process