[Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045)

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
This commit is contained in:
Jee Jee Li
2026-04-11 00:20:20 +08:00
committed by GitHub
parent 8f121f7879
commit ecd1ea1363
14 changed files with 1861 additions and 4 deletions
+1 -3
View File
@@ -233,9 +233,7 @@ class MiniMaxM2Attention(nn.Module):
) -> torch.Tensor:
qkv, _ = self.qkv_proj(hidden_states)
q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
q, k = MiniMaxText01RMSNormTP.forward_qk(
self.q_norm, self.k_norm, q.contiguous(), k.contiguous()
)
q, k = MiniMaxText01RMSNormTP.forward_qk(self.q_norm, self.k_norm, q, k)
q, k = self.rotary_emb(positions, q, k)
attn_output = self.attn(q, k, v)
output, _ = self.o_proj(attn_output)