Data is the cornerstone of artificial intelligence. The efficiency of data acquisition, exchange, and application directly impacts the advances in technologies and applications. Over the long history of AI, a vast quantity of data sets have been developed and distributed. However, these datasets are defined in very different forms, which incurs significant overhead when it comes to exchange, integration, and utilization -- it is often the case that one needs to develop a new customized tool or script in order to incorporate a new dataset into a workflow.

To overcome such difficulties, we develop Data Set Description Language (DSDL). More details please visit our official documents, dsdl datasets can be downloaded from our platform OpenDataLab.

• install dsdl:

  install by pip:

  pip install dsdl
  

  install by source code:

  git clone https://github.com/opendatalab/dsdl-sdk.git -b schema-dsdl
  cd dsdl-sdk
  python setup.py install
  

• install mmdet and pytorch: please refer this installation documents.

• train:

  • using single gpu:

  python tools/train.py {config_file}
  

  • using slurm:

  ./tools/slurm_train.sh {partition} {job_name} {config_file} {work_dir} {gpu_nums}
  

• detection task:

  Datasets	Model	box AP	Config
  VOC07+12	model	80.3*	config
  COCO	model	37.4	config
  Objects365	model	19.8	config
  OpenImages	model	59.9*	config

  *: box AP in voc metric and openimages metric, actually means AP_50.

• instance segmentation task:

  Datasets	Model	box AP	mask AP	Config
  COCO	model	38.1	34.7	config