Chranos/enginex-mlu370-vllm
1
0
Fork 0
forked from EngineX-Cambricon/enginex-mlu370-vllm
Code Pull Requests Activity

add deepseekv3 and llama4

Browse Source
This commit is contained in:
Chranos
2026-02-11 15:39:35 +08:00
parent ed6a2aff91
commit f3a4d10195
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
vllm-v0.6.2/vllm/worker/mlu_worker.py
View File

@@ -60,6 +60,7 @@ class MLUWorker(Worker):
speculative_config = self.speculative_config speculative_config = self.speculative_config
model_config = self.model_config model_config = self.model_config
is_mtp = (speculative_config is not None is_mtp = (speculative_config is not None
and model_config.task != "draft"
and getattr( and getattr(
speculative_config.draft_model_config.hf_config, speculative_config.draft_model_config.hf_config,
"model_type", None) == "deepseek_mtp") "model_type", None) == "deepseek_mtp")