mmdeploy/configs/mmocr/text-detection/text-detection_sdk_dynamic.py

9 lines
260 B
Python
Raw Normal View History

_base_ = ['./text-detection_dynamic.py', '../../_base_/backends/sdk.py']
codebase_config = dict(model_type='sdk')
backend_config = dict(pipeline=[
dict(type='LoadImageFromFile'),
dict(type='PackTextDetInputs', meta_keys=['img_path', 'ori_shape'])
])