The barebones PyTorch backend should be in in the next PyTensor release and subsequent PyMC.
I think the best way to test it out and gather info on what Ops are most needed is to give access from nutpie, as it will:
1) Feel like experimental on experimental
2) Be tested by more code-savy people
The barebones PyTorch backend should be in in the next PyTensor release and subsequent PyMC.
I think the best way to test it out and gather info on what Ops are most needed is to give access from nutpie, as it will: 1) Feel like experimental on experimental 2) Be tested by more code-savy people
What do you think?