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

[New Model]: google/embeddinggemma-300m (#24318)

Browse Source 
Signed-off-by: wang.yuqi <noooop@126.com>
This commit is contained in:
wang.yuqi
2025-09-06 13:58:36 +08:00
committed by GitHub
parent 53b19ccdd5
commit 6d6c6b05d3
9 changed files with 73 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
vllm/model_executor/models/config.py
View File

@@ -24,6 +24,14 @@ class VerifyAndUpdateConfig:
raise NotImplementedError
class Gemma3TextModelConfig:
@staticmethod
def verify_and_update_config(vllm_config: "VllmConfig") -> None:
hf_config = vllm_config.model_config.hf_config
hf_config.is_causal = not hf_config.use_bidirectional_attention
class GteNewModelConfig(VerifyAndUpdateConfig):
@staticmethod
@@ -409,6 +417,7 @@ MODELS_CONFIG_MAP: dict[str, type[VerifyAndUpdateConfig]] = {
"GteModel": SnowflakeGteNewModelConfig,
"GteNewModel": GteNewModelConfig,
"GteNewForSequenceClassification": GteNewModelConfig,
"Gemma3TextModel": Gemma3TextModelConfig,
"NomicBertModel": NomicBertModelConfig,
"Qwen2ForProcessRewardModel": Qwen2ForProcessRewardModelConfig,
"Qwen2ForRewardModel": Qwen2ForRewardModelConfig,