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

[Core] Gate prompt_embeds behind a feature flag (#17607)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-05-04 00:19:20 +08:00
committed by GitHub
parent a92842454c
commit 887d7af882
8 changed files with 84 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/engine/arg_utils.py
View File

@@ -234,6 +234,7 @@ class EngineArgs:
hf_config_path: Optional[str] = ModelConfig.hf_config_path
task: TaskOption = ModelConfig.task
skip_tokenizer_init: bool = ModelConfig.skip_tokenizer_init
enable_prompt_embeds: bool = ModelConfig.enable_prompt_embeds
tokenizer_mode: TokenizerMode = ModelConfig.tokenizer_mode
trust_remote_code: bool = ModelConfig.trust_remote_code
allowed_local_media_path: str = ModelConfig.allowed_local_media_path
@@ -445,6 +446,8 @@ class EngineArgs:
**model_kwargs["disable_cascade_attn"])
model_group.add_argument("--skip-tokenizer-init",
**model_kwargs["skip_tokenizer_init"])
model_group.add_argument("--enable-prompt-embeds",
**model_kwargs["enable_prompt_embeds"])
model_group.add_argument("--served-model-name",
**model_kwargs["served_model_name"])
# This one is a special case because it is the
@@ -874,6 +877,7 @@ class EngineArgs:
disable_sliding_window=self.disable_sliding_window,
disable_cascade_attn=self.disable_cascade_attn,
skip_tokenizer_init=self.skip_tokenizer_init,
enable_prompt_embeds=self.enable_prompt_embeds,
served_model_name=self.served_model_name,
limit_mm_per_prompt=self.limit_mm_per_prompt,
use_async_output_proc=not self.disable_async_output_proc,