Xuanzhao Gao

promote_type stack overflow

3

In ```Julia @inline function get_size_dict(ixs, xs, size_info=nothing) LT = promote_type(eltype.(ixs)...) return get_size_dict!(ixs, xs, size_info===nothing ? Dict{LT,Int}() : size_info) end ``` the function `promote_type` is directly used and I found when...

Hello, I want to upload so package for GPU, which needs `nvcc` to build. Could you please tell me how can that be done? It seems that the sandbox offered...

See https://github.com/Cjkkkk/CUDA_gemm/issues/6.

Error: `corrupted size vs. prev_size`

3

Meet error when contracting of tropical tensor network: ```Julia julia> using TropicalGEMM, TensorInference, JLD, OMEinsum julia> tn = load("./ObjectDetection_32.jld")["tn"] TensorNetworkModel{Int64, SlicedEinsum{Int64, DynamicNestedEinsum{Int64}}, Array{Float64}} variables: 1, 2, 3, 4, 5, 6,...

Plan to add a backend using tree decomposition, which will be provided in TamakiTreeWidth.jl. A data convertor between OMEinsumContractionOrder.jl and TamakiTreeWidth.jl, to convent tensor networks as graphs and to convert...

Adding OMEinsumContractionOrders.jl as a backend of TensorOperations.jl for finding the optimal contraction order

9

An extension `OMEinsumContractionOrdersExt` has been added, which provides `OMEinsumContractionOrders.jl` as a backend of optimizing the contraction order. We modified the `@tensor` marco and added a keyword `opt_algorithm` for selecting a...