Skip to content

Commit 38db258

Browse files
authored
feat: add jinaai/jina-embeddings-v4-vllm-retrieval (#144)
* feat: add jinaai/jina-embeddings-v4-vllm-retrieval * doc: update doc for jinaai/jina-embeddings-v4-vllm-retrieval * chore: code cleanup
1 parent 4329fce commit 38db258

File tree

2 files changed

+30
-0
lines changed

2 files changed

+30
-0
lines changed

docs/en/supported_models.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -71,6 +71,7 @@
7171
| bge-base-en-v1.5 | bge | embedding | g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
7272
| bge-m3 | bge | embedding | g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
7373
| jina-embeddings-v3 | jina | embedding | g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
74+
| jina-embeddings-v4 | jina | embedding | g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
7475
| bge-reranker-v2-m3 | bge | rerank | g4dn.2xlarge,g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
7576
| bge-reranker-large | bge | rerank | g4dn.2xlarge,g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||
7677
| jina-reranker-v2-base-multilingual | jina | rerank | g5.xlarge,g5.2xlarge,g5.4xlarge,g5.8xlarge,g5.16xlarge | sagemaker_realtime,ecs ||

src/emd/models/embeddings/jina.py

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -53,3 +53,32 @@
5353
model_series=JINA_SERIES
5454
)
5555
)
56+
57+
Model.register(
58+
dict(
59+
model_id = "jina-embeddings-v4-vllm-retrieval",
60+
supported_engines=[huggingface_embedding_engine449],
61+
supported_instances=[
62+
g5dxlarge_instance,
63+
g5d2xlarge_instance,
64+
g5d4xlarge_instance,
65+
g5d8xlarge_instance
66+
],
67+
supported_services=[
68+
sagemaker_service,
69+
ecs_service,
70+
local_service
71+
],
72+
supported_frameworks=[
73+
fastapi_framework
74+
],
75+
allow_china_region=True,
76+
huggingface_model_id="jinaai/jina-embeddings-v4-vllm-retrieval",
77+
modelscope_model_id="jinaai/jina-embeddings-v4-vllm-retrieval",
78+
require_huggingface_token=False,
79+
application_scenario="RAG",
80+
model_type=ModelType.EMBEDDING,
81+
description="jina-embeddings-v4 is a universal embedding model for multimodal and multilingual retrieval",
82+
model_series=JINA_SERIES
83+
)
84+
)

0 commit comments

Comments
 (0)