_base_ = ['./text-detection_dynamic.py', '../../_base_/backends/ascend.py'] onnx_config = dict(input_shape=None) backend_config = dict(model_inputs=[ dict( input_shapes=dict(input=[-1, 3, -1, -1]), dynamic_dims=[(1, 640, 640), (4, 640, 640), (1, 1280, 1280)]) ])
