F1 metric

[LinasKo]

Description

This PR adds the F1 score as a metric.

Other changes:

• Removed Internal Data store (dead code)
• Small refactoring of mAP metric.
• One renamed an mAP parameter to ap_per_class
• One new mAP parameter mAP_scores.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

How has this change been tested, please provide a testcase or example of how you tested the change?

I ran it on a small set of tests locally:

• No detections, one image, multiple sets of detections, multiple averaging types, boxes & masks.

TODO: units tests
TODO: Colab
TODO: check docs

Any specific deployment considerations

Docs

• Docs updated? What were the changes:

[LinasKo]

Hey @onuralpszr, if you have some time, could you give this a glance and see if anything weird stands out?