v0.6.5

OMEinsum v0.6.5

Diff since v0.6.4

Merged pull requests:

• manually construct contraction tree and flop counting (#130) (@GiggleLiu)

v0.6.4

OMEinsum v0.6.4

Diff since v0.6.3

Merged pull requests:

• Polish AD and unify some interfaces (#129) (@GiggleLiu)

v0.6.3

OMEinsum v0.6.3

Diff since v0.6.2

Closed issues:

• Maybe a bug: complex CuArray raise error when compiling cu-function (#124)
• Errors when using permutedims on CuArray (#127)

Merged pull requests:

• rm conj! to support hessian (#125) (@GiggleLiu)
• Add a debug statement when we fallback in cueinsum (#126) (@kshyatt)
• fix CUDA high dimensional permutedims (#128) (@GiggleLiu)

v0.6.2

OMEinsum v0.6.2

Diff since v0.6.1

Merged pull requests:

• Fix flatten function for static code (#123) (@GiggleLiu)

v0.6.1

OMEinsum v0.6.1

Diff since v0.6.0

Merged pull requests:

• fix flatten (#122) (@GiggleLiu)

v0.6.0

OMEinsum v0.6.0

Diff since v0.5.0

Merged pull requests:

• Reduce dynamic overheads (#121) (@GiggleLiu)

v0.5.0

OMEinsum v0.5.0

Diff since v0.4.8

Closed issues:

• Benchmarking (#9)
• Increase the performance of dynamic calls (#73)
• switch to BatchedRoutine APIs (#76)

Merged pull requests:

• better time and space complexity computation (#117) (@GiggleLiu)
• Add many-index benchmark (#118) (@kshyatt)
• Seperate dynamic eincode and static eincode (#119) (@GiggleLiu)

v0.4.8

OMEinsum v0.4.8

Diff since v0.4.7

Merged pull requests:

• Fix 1.5 compatibility (#116) (@GiggleLiu)

v0.4.7

OMEinsum v0.4.7

Diff since v0.4.6

v0.4.6

OMEinsum v0.4.6

Diff since v0.4.5

Closed issues:

• Suppress "OMEinsum: YOU FIND CUDA!" message (#114)

Merged pull requests:

• dynamic nested einsum (#112) (@GiggleLiu)
• update deps (#113) (@GiggleLiu)
• use @info to print CUDA loading message. (#115) (@GiggleLiu)