I believe I’ve found a solution to issue #69519. Which is an issue regarding a lack of implementation of the histogram method for CUDA tensors. There is no function being dispatched to using CUDA in aten/src/ATen/native/native_functions.yaml
The related cu file for a related histc file is found in SummaryOps.cu
I would appreciate some feedback on contributing to Pytorch for this specific issue. I will probably write an RFC on the issue when I have a clearer picture of my solution. But before I start I want to know if this is considered too large. Or even just information relating to the level of effort for this.
(Also excuse me if this is the wrong category label, would appreciate feedback on any meta contributing things)