Hello, we need some clarifications about the final project,
- In the “Deep learning project-submission” post in the Moodle, it says “Choose the classification/segmentation task”, is that correct? Or is it supposed to be “object-detection” instead of “classification”?
- Are we required to work on a pre-trained model or can we train a model from scratch?
- Can the 2 datasets come from the same source?(E.g. 2 datasets from COCO)
- Can we train the model on a slice from each dataset or are we required to use the whole datasets?
- How many objects/classes is the model supposed to detect/classify?