opt_einsum_fx

opt_einsum_fx copied to clipboard

It could be useful to provide some way of marking a Tensor as a scalar and adding it to the propagation logic.

Linux-cpp-lisp

Consider: - [ ] `torch.reshape` - [ ] `torch.cross` - [ ] `torch.dot` - [ ] `torch.transpose` - [ ] `torch.nn.functional.bilinear`

Linux-cpp-lisp

Make a transformer that transforms `x.transpose(1, 2)` into `torch.einsum('abc...->acb...')` in order to then have these operations fused with the rest of the einsums

mariogeiger

When accumulating scalar constants at graph optimization time, arbitrary precision arithmetic should be used to ensure the best result.

Linux-cpp-lisp

Accumulate the theoretical speedup, scaling factor, and intermediate sizes across einsums processed in a graph and somehow report them.

Linux-cpp-lisp