Paul Fultz II
Paul Fultz II
This implements a faster GPU topk. * Update the ref version of topk to take a parameter for the indices, and also updated to handle any layout. * Added a...
Adds a jenkins job that will run the unit tests with debug mode with libstdc++. To fix the debug issues I wrapped the instruction_ref in a class(actually reused the class...
This exposes the `outputs` and `get_literal` methods on the instruction and exposes the `replace_return` on the module.
This allows us to easily extract subgraphs using the driver directly.