frozenleaves/vllm
1
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2025-10-20 14:53:52 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix imports for MoE on CPU (#15841)

Browse Source 
Signed-off-by: Thien Tran <gau.nernst@yahoo.com.sg>
This commit is contained in:
Thien Tran
2025-04-02 11:33:55 +08:00
committed by GitHub
parent 6efb195a6e
commit 2039c6305b
1 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/model_executor/layers/fused_moe/rocm_aiter_fused_moe.py
View File

@ -4,8 +4,6 @@ from typing import List, Optional
import torch
import vllm.envs as envs
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
per_token_group_quant_fp8)
from vllm.platforms import current_platform
@ -38,6 +36,9 @@ def rocm_aiter_fused_experts(
import aiter as rocm_aiter
import aiter.fused_moe_bf16_asm as rocm_aiter_asm_fmoe
from vllm.model_executor.layers.quantization.utils.fp8_utils import (
per_token_group_quant_fp8)
if envs.VLLM_ROCM_USE_AITER_FP8_BLOCK_SCALED_MOE and use_fp8_w8a8:
assert w1_scale is not None
assert w2_scale is not None