microsoft / BitBLAS

BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.
MIT License
190 stars 21 forks source link

Fix gpu model missing from tvm target remap #61

Closed Qubitium closed 2 days ago

Qubitium commented 2 days ago

A model was missing in the previously merged nvidia oem a100 remap. Refractored so the code is not static but a dictionary lookup for better future additions. Ideally, this code should be part of tvm and not downstream consumer of tvm like bitlbas.