-
Notifications
You must be signed in to change notification settings - Fork 149
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support multiple types for Kompute Tensors #2
Comments
Currently refactored to make float the default - needs further exploration on best way to provide typing. The options are to make the Tensor a template, or to add the type as a member variables. It will be important to assess the advantages/disadvantages of each. |
@alexander-g I was looking that the vkjax does seem to somewhat support multiple dtypes, and I was currently assessing whether it would be worth adding the ability to support multiple types, for the Kompute tensors - is this something that would add a significant amount of value? Or would it be the case that most scientific compute can be done primarily through float32 types throughout? |
Currently only uint32_t is supported but looking to support further types. This will also require evaluating how these are managed by "Algorithms" and "ParameterGroups"
The text was updated successfully, but these errors were encountered: