Closed jcmgray closed 3 years ago
This PR:
memory_limit
'dp'
tensordot
isinstance
infer_backend
Merging #154 into master will increase coverage by 0.00%. The diff coverage is 100.00%.
0.00%
100.00%
Description
This PR:
memory_limit
meant the'dp'
optimizer would search forever (#153)tensordot
axes in a canonical order (the order they appear on the first operand), so that performance shouldn't change unexpectedly (#143)isinstance
call toinfer_backend
(is cleaner? and fixes a rare bug when mixing inputs for e.g. jax compiled contraction)Status