[Model] Enable LoRA support for RobertaForSequenceClassification (#58884)
Signed-off-by: jz-yolo <jz-yolo@users.noreply.github.com> Signed-off-by: Jane Zhu <jane.zhu@slack-corp.com> Co-authored-by: jz-yolo <jz-yolo@users.noreply.github.com>
This commit is contained in:
@@ -50,8 +50,8 @@ The three supported scoring functions are as illustrated in the figure below.
|
||||
| `ModernBertForSequenceClassification` | ModernBERT-based | `Alibaba-NLP/gte-reranker-modernbert-base`, etc. | N/A | | |
|
||||
| `Qwen2ForSequenceClassification`<sup>C</sup> | Qwen2-based | `mixedbread-ai/mxbai-rerank-base-v2`(see note), etc. | [mxbai_rerank_v2.jinja](../../../examples/pooling/score/template/mxbai_rerank_v2.jinja) | ✅︎ | ✅︎ |
|
||||
| `Qwen3ForSequenceClassification`<sup>C</sup> | Qwen3-based | `tomaarsen/Qwen3-Reranker-0.6B-seq-cls`, `Qwen/Qwen3-Reranker-0.6B`(see note), etc. | [qwen3_reranker.jinja](../../../examples/pooling/score/template/qwen3_reranker.jinja) | ✅︎ | ✅︎ |
|
||||
| `RobertaForSequenceClassification` | RoBERTa-based | `cross-encoder/quora-roberta-base`, etc. | N/A | | |
|
||||
| `XLMRobertaForSequenceClassification` | XLM-RoBERTa-based | `BAAI/bge-reranker-v2-m3`, etc. | N/A | | |
|
||||
| `RobertaForSequenceClassification` | RoBERTa-based | `cross-encoder/quora-roberta-base`, etc. | N/A | ✅︎ | |
|
||||
| `XLMRobertaForSequenceClassification` | XLM-RoBERTa-based | `BAAI/bge-reranker-v2-m3`, etc. | N/A | ✅︎ | |
|
||||
| `*Model`<sup>C</sup>, `*ForCausalLM`<sup>C</sup>, etc. | Generative models | N/A | N/A | \* | \* |
|
||||
|
||||
<sup>C</sup> Automatically converted into a classification model via `--convert classify`. ([details](./README.md#model-conversion))
|
||||
|
||||
@@ -41,7 +41,7 @@ from vllm.model_executor.models.utils import (
|
||||
from vllm.sequence import IntermediateTensors
|
||||
|
||||
from .bert_with_rope import BertWithRope, JinaRobertaModel
|
||||
from .interfaces import SupportsCrossEncoding
|
||||
from .interfaces import SupportsCrossEncoding, SupportsLoRA
|
||||
from .interfaces_base import attn_type, default_pooling_type
|
||||
|
||||
|
||||
@@ -266,7 +266,7 @@ class BgeM3EmbeddingModel(RobertaEmbeddingModel):
|
||||
|
||||
|
||||
@default_pooling_type(seq_pooling_type="CLS")
|
||||
class RobertaForSequenceClassification(nn.Module, SupportsCrossEncoding):
|
||||
class RobertaForSequenceClassification(nn.Module, SupportsCrossEncoding, SupportsLoRA):
|
||||
"""A model that uses Roberta to provide embedding functionalities.
|
||||
|
||||
This class encapsulates the BertModel and provides an interface for
|
||||
@@ -279,6 +279,7 @@ class RobertaForSequenceClassification(nn.Module, SupportsCrossEncoding):
|
||||
"""
|
||||
|
||||
is_pooling_model = True
|
||||
packed_modules_mapping = {"qkv_proj": ["query", "key", "value"]}
|
||||
jina_to_vllm_mapper = WeightsMapper(
|
||||
orig_to_new_substr={
|
||||
"emb_ln": "embeddings.LayerNorm",
|
||||
@@ -295,7 +296,8 @@ class RobertaForSequenceClassification(nn.Module, SupportsCrossEncoding):
|
||||
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
|
||||
super().__init__()
|
||||
config = vllm_config.model_config.hf_config
|
||||
self.padding_idx: int = vllm_config.model_config.hf_config.pad_token_id
|
||||
self.config = config
|
||||
self.padding_idx: int = config.pad_token_id
|
||||
|
||||
self.num_labels = config.num_labels
|
||||
self.roberta = BertModel(
|
||||
|
||||
Reference in New Issue
Block a user