efficientdet_d2¶

lucid.models.efficientdet_d2(num_classes: int = 80, **kwargs) → EfficientDet¶

The efficientdet_d2 function returns an instance of the EfficientDet model, configured with compound coefficient 2, offering improved accuracy through deeper layers and increased input resolution.

Total Parameters: 6,457,434

Function Signature¶

def efficientdet_d2(num_classes: int = 80, **kwargs) -> EfficientDet

Parameters¶

num_classes (int, default=80): Number of target object classes to detect.
kwargs: Additional keyword arguments passed to the EfficientDet constructor.

Returns¶

EfficientDet: An instance of the EfficientDet-D2 model with EfficientNet-B2 backbone and scaled BiFPN depth.

Example Usage¶

>>> from lucid.models import efficientdet_d2
>>> model = efficientdet_d2(num_classes=80)
>>> print(model)

>>> x = lucid.rand(2, 3, 768, 768)
>>> out = model(x)
>>> for o in out:
...     print(o.shape)