mirror of
https://github.com/wassname/vllm.git
synced 2026-08-06 13:40:27 +08:00
[Speculative decoding 7/9] Speculative decoding end-to-end correctness tests. (#3951)
This commit is contained in:
@@ -83,6 +83,7 @@ class GPUExecutor(ExecutorBase):
|
||||
scheduler_config=self.scheduler_config,
|
||||
device_config=self.device_config,
|
||||
cache_config=self.cache_config,
|
||||
# TODO allow draft-model specific load config.
|
||||
load_config=self.load_config,
|
||||
local_rank=0,
|
||||
rank=0,
|
||||
|
||||
Reference in New Issue
Block a user