There was support for a num_beams parameter before but now the sampling parameters has no such parameter. Can support for beam search be restored?
Activity
github-actions commentedon Jul 25, 2024
This issue has been automatically closed due to inactivity. Please feel free to reopen it if needed.
add: w4a8 group gemm (sgl-project#353)
Revert "add: w4a8 group gemm (sgl-project#353)" (sgl-project#360)
refactor: move all pallas kernels to 'python/sgl_jax/kernels' folder (s…