diff --git a/vllm/model_executor/models/voyage.py b/vllm/model_executor/models/voyage.py index d131ffff4b..9f4885a0f9 100644 --- a/vllm/model_executor/models/voyage.py +++ b/vllm/model_executor/models/voyage.py @@ -8,6 +8,7 @@ import torch import torch.nn as nn from vllm.config import VllmConfig +from vllm.model_executor.models.interfaces import SupportsLoRA from vllm.model_executor.models.qwen3 import Qwen3Model from vllm.model_executor.models.utils import AutoWeightsLoader, maybe_prefix from vllm.sequence import IntermediateTensors @@ -15,7 +16,7 @@ from vllm.sequence import IntermediateTensors WeightItem = tuple[str, torch.Tensor] -class VoyageQwen3BidirectionalEmbedModel(nn.Module): +class VoyageQwen3BidirectionalEmbedModel(nn.Module, SupportsLoRA): """Qwen3Model + Voyage embedding head + bidirectional attention. Checkpoint conventions (HF): @@ -30,6 +31,11 @@ class VoyageQwen3BidirectionalEmbedModel(nn.Module): - No "model." prefix """ + packed_modules_mapping = { + "qkv_proj": ["q_proj", "k_proj", "v_proj"], + "gate_up_proj": ["gate_proj", "up_proj"], + } + def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""): super().__init__() self.config = vllm_config.model_config.hf_config