get_efficientdet

hybrid_learning.experimentation.model_registry.fuzzy_exp_models.get_efficientdet(img_size, model_key='tf_efficientdet_d1', effdet_wrapper='predict', **_)[source]

Obtain a standard EfficientDet model.

Parameters

img_size – image size in (height, width); each dimension must be divisible by 2**max_level (usually 128)
model_key – the EfficientDet variant specifier; for options see https://github.com/rwightman/efficientdet-pytorch/blob/75e16c2f/effdet/config/model_config.py#L82
effdet_wrapper – whether to wrap the EfficientDet model for prediction ('prediction', add NMS), for training ('train'), or not at all ('')

Return type

Module