[Docs] update docs (#2573)

open-mmlab · Jul 26, 2023 · a802a3d · a802a3d
1 parent e8bf5a7
commit a802a3d
Show file tree

Hide file tree

Showing 6 changed files with 46 additions and 46 deletions.
diff --git a/docs/en/advanced_guides/customize_datasets.md b/docs/en/advanced_guides/customize_datasets.md
@@ -213,7 +213,7 @@ The following dataset wrappers are supported in [MMEngine](https://github.com/op
 
 ### CombinedDataset
 
-MMPose provides `CombinedDataset` to combine multiple datasets with different annotations. A combined dataset can be defined in config files as:
+MMPose provides [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) to combine multiple datasets with different annotations. A combined dataset can be defined in config files as:
 
 ```python
 dataset_1 = dict(
@@ -250,6 +250,6 @@ combined_dataset = dict(
 
 - **MetaInfo of combined dataset** determines the annotation format. Either metainfo of a sub-dataset or a customed dataset metainfo is valid here. To custom a dataset metainfo, please refer to [Create a custom dataset_info config file for the dataset](#create-a-custom-datasetinfo-config-file-for-the-dataset).
 
-- **Converter transforms of sub-datasets** are applied when there exist mismatches of annotation format between sub-datasets and the combined dataset. For example, the number and order of keypoints might be different in the combined dataset and the sub-datasets. Then `KeypointConverter` can be used to unify the keypoints number and order.
+- **Converter transforms of sub-datasets** are applied when there exist mismatches of annotation format between sub-datasets and the combined dataset. For example, the number and order of keypoints might be different in the combined dataset and the sub-datasets. Then [KeypointConverter](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/transforms/converting.py#L11) can be used to unify the keypoints number and order.
 
-- More details about `CombinedDataset` and `KeypointConverter` can be found in Advanced Guides-[Training with Mixed Datasets](../user_guides/mixed_datasets.md).
+- More details about [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) and [KeypointConverter](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/transforms/converting.py#L11) can be found in [Advanced Guides - Training with Mixed Datasets](../user_guides/mixed_datasets.md).
diff --git a/docs/en/user_guides/inference.md b/docs/en/user_guides/inference.md
@@ -9,11 +9,11 @@ In MMPose, a model is defined by a configuration file, while its pre-existing pa
 
 ## Inferencer: a Unified Inference Interface
 
-MMPose offers a comprehensive API for inference, known as `MMPoseInferencer`. This API enables users to perform inference on both images and videos using all the models supported by MMPose. Furthermore, the API provides automatic visualization of inference results and allows for the convenient saving of predictions.
+MMPose offers a comprehensive API for inference, known as [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24). This API enables users to perform inference on both images and videos using all the models supported by MMPose. Furthermore, the API provides automatic visualization of inference results and allows for the convenient saving of predictions.
 
 ### Basic Usage
 
-The `MMPoseInferencer` can be used in any Python program to perform pose estimation. Below is an example of inference on a given image using the pre-trained human pose estimator within the Python shell.
+The [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24) can be used in any Python program to perform pose estimation. Below is an example of inference on a given image using the pre-trained human pose estimator within the Python shell.
 
 ```python
 from mmpose.apis import MMPoseInferencer
@@ -80,7 +80,7 @@ python demo/inferencer_demo.py 'tests/data/coco/000000000785.jpg' \
     --pose2d 'human' --show --pred-out-dir 'predictions'
 ```
 
-The predictions will be save in `predictions/000000000785.json`. The argument names correspond with the `MMPoseInferencer`, which serves as an API.
+The predictions will be save in `predictions/000000000785.json`. The argument names correspond with the [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24), which serves as an API.
 
 The inferencer is capable of processing a range of input types, which includes the following:
 
@@ -219,7 +219,7 @@ result = next(result_generator)
 
 ### Arguments of Inferencer
 
-The `MMPoseInferencer` offers a variety of arguments for customizing pose estimation, visualization, and saving predictions. Below is a list of the arguments available when initializing the inferencer and their descriptions:
+The [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24) offers a variety of arguments for customizing pose estimation, visualization, and saving predictions. Below is a list of the arguments available when initializing the inferencer and their descriptions:
 
 | Argument         | Description                                                                                                      |
 | ---------------- | ---------------------------------------------------------------------------------------------------------------- |
@@ -233,7 +233,7 @@ The `MMPoseInferencer` offers a variety of arguments for customizing pose estima
 | `device`         | The device to perform the inference. If left `None`, the Inferencer will select the most suitable one.           |
 | `scope`          | The namespace where the model modules are defined.                                                               |
 
-The inferencer is designed for both visualization and saving predictions. The table below presents the list of arguments available when using the `MMPoseInferencer` for inference, along with their compatibility with 2D and 3D inferencing:
+The inferencer is designed for both visualization and saving predictions. The table below presents the list of arguments available when using the [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24) for inference, along with their compatibility with 2D and 3D inferencing:
 
 | Argument                  | Description                                                                                                                                                       | 2D  | 3D  |
 | ------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | --- | --- |
@@ -258,7 +258,7 @@ The inferencer is designed for both visualization and saving predictions. The ta
 
 ### Model Alias
 
-The MMPose library has predefined aliases for several frequently used models. These aliases can be utilized as a shortcut when initializing the `MMPoseInferencer`, as an alternative to providing the full model configuration name. Here are the available 2D model aliases and their corresponding configuration names:
+The MMPose library has predefined aliases for several frequently used models. These aliases can be utilized as a shortcut when initializing the [MMPoseInferencer](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/apis/inferencers/mmpose_inferencer.py#L24), as an alternative to providing the full model configuration name. Here are the available 2D model aliases and their corresponding configuration names:
 
 | Alias     | Configuration Name                                 | Task                            | Pose Estimator | Detector            |
 | --------- | -------------------------------------------------- | ------------------------------- | -------------- | ------------------- |

diff --git a/docs/en/user_guides/mixed_datasets.md b/docs/en/user_guides/mixed_datasets.md
@@ -1,10 +1,10 @@
 # Use Mixed Datasets for Training
 
-MMPose offers a convenient and versatile solution for training with mixed datasets through its `CombinedDataset` tool. Acting as a wrapper, it allows for the inclusion of multiple datasets and seamlessly reads and converts data from varying sources into a unified format for model training. The data processing pipeline utilizing `CombinedDataset` is illustrated in the following figure.
+MMPose offers a convenient and versatile solution for training with mixed datasets through its [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) tool. Acting as a wrapper, it allows for the inclusion of multiple datasets and seamlessly reads and converts data from varying sources into a unified format for model training. The data processing pipeline utilizing [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) is illustrated in the following figure.
 
 ![combined_dataset_pipeline](https://user-images.githubusercontent.com/26127467/223333154-fb88e511-810a-423c-b755-c791d296bc43.jpg)
 
-The following section will provide a detailed description of how to configure `CombinedDataset` with an example that combines the COCO and AI Challenger (AIC) datasets.
+The following section will provide a detailed description of how to configure [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) with an example that combines the COCO and AI Challenger (AIC) datasets.
 
 ## COCO & AIC example
 
@@ -39,7 +39,7 @@ dataset_coco = dict(
 )
 ```
 
-For AIC dataset, the order of the keypoints needs to be transformed. MMPose provides a `KeypointConverter` transform to achieve this. Here's an example of how to configure the AIC sub dataset:
+For AIC dataset, the order of the keypoints needs to be transformed. MMPose provides a [KeypointConverter](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/transforms/converting.py#L11) transform to achieve this. Here's an example of how to configure the AIC sub dataset:
 
 ```python
 dataset_aic = dict(
@@ -70,9 +70,9 @@ dataset_aic = dict(
 )
 ```
 
-By using the `KeypointConverter`, the indices of keypoints with indices 0 to 11 will be transformed to corresponding indices among 5 to 16. Meanwhile, the keypoints with indices 12 and 13 will be removed. For the target keypoints with indices 0 to 4, which are not defined in the `mapping` argument, they will be set as invisible and won't be used in training.
+By using the [KeypointConverter](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/transforms/converting.py#L11), the indices of keypoints with indices 0 to 11 will be transformed to corresponding indices among 5 to 16. Meanwhile, the keypoints with indices 12 and 13 will be removed. For the target keypoints with indices 0 to 4, which are not defined in the `mapping` argument, they will be set as invisible and won't be used in training.
 
-Once the sub datasets are configured, the `CombinedDataset` wrapper can be defined as follows:
+Once the sub datasets are configured, the [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) wrapper can be defined as follows:
 
 ```python
 dataset = dict(
@@ -100,7 +100,7 @@ The previously mentioned method discards some annotations in the AIC dataset. If
 
 <img src="https://user-images.githubusercontent.com/26127467/223356617-075e0ab1-0ed3-426d-bc88-4f16be93f0ba.png" height="200px" alt><br>
 
-In this scenario, both COCO and AIC datasets need to adjust the keypoint indices using `KeypointConverter`:
+In this scenario, both COCO and AIC datasets need to adjust the keypoint indices using [KeypointConverter](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/transforms/converting.py#L11):
 
 ```python
 dataset_coco = dict(
@@ -172,7 +172,7 @@ When training with mixed datasets, users often encounter the problem of inconsis
 
 ### Adjust the sampling ratio of each sub dataset
 
-In `CombinedDataset`, we provide the `sample_ratio_factor` argument to adjust the sampling ratio of each sub dataset.
+In [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15), we provide the `sample_ratio_factor` argument to adjust the sampling ratio of each sub dataset.
 
 For example:
 

diff --git a/docs/zh_cn/advanced_guides/customize_datasets.md b/docs/zh_cn/advanced_guides/customize_datasets.md
@@ -217,14 +217,14 @@ test_dataloader = dict(
 
 ## 数据集封装
 
-目前 [MMEngine](https://github.com/open-mmlab/mmengine) 支持以下数据集封装：
+在 MMPose 中，支持使用 MMPose 实现的数据集封装和 [MMEngine](https://github.com/open-mmlab/mmengine) 实现的数据集封装。目前 [MMEngine](https://github.com/open-mmlab/mmengine) 支持以下数据集封装：
 
 - [ConcatDataset](https://mmengine.readthedocs.io/zh_CN/latest/advanced_tutorials/basedataset.html#concatdataset)
 - [RepeatDataset](https://mmengine.readthedocs.io/zh_CN/latest/advanced_tutorials/basedataset.html#repeatdataset)
 
 ### CombinedDataset
 
-MMPose 提供了一个 `CombinedDataset` 类，它可以将多个数据集封装成一个数据集。它的使用方法如下：
+MMPose 提供了一个 [CombinedDataset](https://github.com/open-mmlab/mmpose/blob/dev-1.x/mmpose/datasets/dataset_wrappers.py#L15) 类，它可以将多个数据集封装成一个数据集。它的使用方法如下：
 
 ```python
 dataset_1 = dict(