Enable prediction using Detectree2 model

[amrithasp02]

This PR focuses on including Detectree2 model into the framework. All steps required to perform batch predictions using the dataloader have been included.

Main components include:

1. Detectree2Detector in detector.py
2. Detectree2Module in models.py - this class currently only deals with setting up the model config. When we implement the training portion, it will be used in a way similar to DeepForestModule.
3. detectree2_detections.ipynb

[russelldj]

Also, what do you think about adding this predictor as an option in the predict detections example? You could provide the config constants for both in cell 2 as well as another one to select which model to use. And then in cell 5 create the appropriate detector.

[russelldj]

From the discussion, we need to make sure that this uses the GPU.