[M3] Enable FP8 sparse GQA (#45744)
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg> Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>
This commit is contained in:
@@ -471,7 +471,8 @@ STABLE_TORCH_LIBRARY_FRAGMENT(_C, ops) {
|
||||
"int num_index_heads, "
|
||||
"Tensor? slot_mapping, Tensor? index_slot_mapping, "
|
||||
"Tensor!? kv_cache, Tensor!? index_cache, "
|
||||
"int block_size, Tensor!? q_out, Tensor!? index_q_out) -> ()");
|
||||
"int block_size, Tensor!? q_out, Tensor!? index_q_out, "
|
||||
"str kv_cache_dtype) -> ()");
|
||||
|
||||
// Apply repetition penalties to logits in-place.
|
||||
ops.def(
|
||||
|
||||
Reference in New Issue
Block a user