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

[Dev] Bump Version to 0.0.1.dev9 #51

Closed LeiWang1999 closed 3 weeks ago

LeiWang1999 commented 3 weeks ago

This pull request primarily includes updates to the version number, modifications to the CUDA kernel launch string in the legalize_c function, and a change in the test_matmul_torch_forward_fp8e4m3 test function. The version number is updated in two places, the VERSION file and the __init__.py file. The legalize_c function is modified to return early if a certain condition is met. Lastly, the test_matmul_torch_forward_fp8e4m3 function now uses a list instead of a single value for the M parameter.

Version number updates:

Changes in legalize_c function:

Test function modification: