biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Model] Add MoE support for NemotronH (#25863)

Browse Source 
Signed-off-by: Tomer Asida <57313761+tomeras91@users.noreply.github.com>
This commit is contained in:
tomeras91
2025-10-23 13:27:23 +03:00
committed by GitHub
parent 88afa11010
commit 61089465a6
7 changed files with 413 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/layers/fused_moe/fused_moe.py
View File

@@ -1647,6 +1647,7 @@ def fused_experts(
SILU_NO_MUL: str = activation_without_mul("silu")
GELU_NO_MUL: str = activation_without_mul("gelu")
RELU2_NO_MUL: str = activation_without_mul("relu2")
def _get_config_quant_dtype(
@@ -1914,7 +1915,8 @@ def fused_experts_impl(
intermediate_cache2 = F.silu(intermediate_cache1.view(-1, N))
elif activation == GELU_NO_MUL:
intermediate_cache2 = F.gelu(intermediate_cache1.view(-1, N))
elif activation == RELU2_NO_MUL:
intermediate_cache2 = torch.square(F.relu(intermediate_cache1.view(-1, N)))
else:
raise ValueError(f"Unsupported FusedMoe activation: {activation}.")