Add utilities to distributions

[noahfarr]

In pytorch, the following is easily possible:

logits = ...
probs = Categorical(logits=logits)
log_prob = probs.log_prob(value)
entropy = probs.entropy()

but when I want to achieve something similar in MLX, I have to manually calculate the log_prob and entropy. Is it possible to add support for these methods as it makes working with distributions in MLX much more convenient (at least for me)

[awni]

We don't have a distributions package in the same way the PyTorch does and don't currently have a plan yet to add one to MLX core.

We can keep this issue open to gauge priority.. but for now it's not likely something we will implement in the near future. A good short-term strategy is to encourage a third party development of mlx-distributions or something.

[noahfarr]

Thanks for your quick response :)