[Perf] Fix jit compiles at runtime of fla gated delta rule (#25432)
Co-authored-by: Michael Goin <mgoin64@gmail.com> Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
committed by
yewentao256
parent
177c37e960
commit
f52b991db6
@ -40,8 +40,8 @@ def fused_recurrent_gated_delta_rule_fwd_kernel(
|
||||
ssm_state_indices,
|
||||
num_accepted_tokens,
|
||||
scale,
|
||||
N: tl.constexpr, # num of sequences
|
||||
T: tl.constexpr, # num of tokens
|
||||
N: tl.int64, # num of sequences
|
||||
T: tl.int64, # num of tokens
|
||||
B: tl.constexpr,
|
||||
H: tl.constexpr,
|
||||
HV: tl.constexpr,
|
||||
|
||||
Reference in New Issue
Block a user