Kai-Hsin Wu

Results 73 comments of Kai-Hsin Wu

Oh you want to extract the sign but keep the zero as it is? On Wed, Mar 12, 2025, 21:58 Chan-Sing-Hong ***@***.***> wrote: > Yes, but we might need a...

There is. Originally some of the API are design for large Tensor. Meaning that if you don't need to support -1 for indexing like python, the C++ side can utilize...

this is not consistent with numpy

I don't see why you can't reshape after create this

Why not do: ```python uT=cytnx.UniTensor.arange(2*3*4).reshape([2,3,4],rowrank=2).relabel(["a","b","c"]) ```

I think current behavior of reshape will make row rank=0. Which we can just change it to be new_rank/2 On Wed, Sep 4, 2024, 03:28 Pochung Chen ***@***.***> wrote: >...

I am confused. Whats wrong with the iterator? We have a Proxy class that was design to do that. Its carrying a pointer so that the lval and rval get/set...

Another reason for this proxy is that one might access elements from a BlockUniTensor that does not exists, then the proxy allows you to check that.

> And it probably makes sense for UniTensor to be a template as well, right? It needs to contain a bunch of tensor blocks, that all need to be the...

> On the question of templatizing everything, the python interface cannot be templatized, so at some level there needs to be type erasure. I guess that could be a `PyObject`,...