* chore: addingop k categorical accuraracy
* chore: adding top k and in top k
* chore: fixing tests
* chore: y true argmax
* chore: adding sparse top k cat metric
* review coomments
* Added confusion metrics -- still using TF ops
* Fixed structure + tests pass for TF (still need to port to multi-backend)
* Got rid of most tf deps, still a few more to go
* Full removal of TF. Tests pass for both Jax and TF
* Full removal of TF. Tests pass for both Jax and TF
* Formatting
* Formatting
* Review comments
* More review comments + formatting