devMls
commited on
Commit
·
e5a1268
1
Parent(s):
26d4422
add jina new models in jina connector (#3770)
Browse files### What problem does this PR solve?
add new models in jinna connector, to allow use models that support
multilingual models
### Type of change
- [X] Other (please describe): new connectors no breaking change
- conf/llm_factories.json +36 -0
- rag/llm/embedding_model.py +1 -1
- rag/llm/rerank_model.py +1 -1
conf/llm_factories.json
CHANGED
|
@@ -525,6 +525,18 @@
|
|
| 525 |
"tags": "TEXT EMBEDDING",
|
| 526 |
"max_tokens": 8196,
|
| 527 |
"model_type": "embedding"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 528 |
}
|
| 529 |
]
|
| 530 |
},
|
|
@@ -2432,6 +2444,18 @@
|
|
| 2432 |
"max_tokens": 4000,
|
| 2433 |
"model_type": "embedding"
|
| 2434 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2435 |
{
|
| 2436 |
"llm_name": "rerank-1",
|
| 2437 |
"tags": "RE-RANK, 8000",
|
|
@@ -2443,6 +2467,18 @@
|
|
| 2443 |
"tags": "RE-RANK, 4000",
|
| 2444 |
"max_tokens": 4000,
|
| 2445 |
"model_type": "rerank"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2446 |
}
|
| 2447 |
]
|
| 2448 |
},
|
|
|
|
| 525 |
"tags": "TEXT EMBEDDING",
|
| 526 |
"max_tokens": 8196,
|
| 527 |
"model_type": "embedding"
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"llm_name": "jina-reranker-v2-base-multilingual",
|
| 531 |
+
"tags": "RE-RANK,8k",
|
| 532 |
+
"max_tokens": 8196,
|
| 533 |
+
"model_type": "rerank"
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"llm_name": "jina-embeddings-v3",
|
| 537 |
+
"tags": "TEXT EMBEDDING",
|
| 538 |
+
"max_tokens": 8196,
|
| 539 |
+
"model_type": "embedding"
|
| 540 |
}
|
| 541 |
]
|
| 542 |
},
|
|
|
|
| 2444 |
"max_tokens": 4000,
|
| 2445 |
"model_type": "embedding"
|
| 2446 |
},
|
| 2447 |
+
{
|
| 2448 |
+
"llm_name": "voyage-3",
|
| 2449 |
+
"tags": "TEXT EMBEDDING,32000",
|
| 2450 |
+
"max_tokens": 32000,
|
| 2451 |
+
"model_type": "embedding"
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"llm_name": "voyage-3-lite",
|
| 2455 |
+
"tags": "TEXT EMBEDDING,32000",
|
| 2456 |
+
"max_tokens": 32000,
|
| 2457 |
+
"model_type": "embedding"
|
| 2458 |
+
},
|
| 2459 |
{
|
| 2460 |
"llm_name": "rerank-1",
|
| 2461 |
"tags": "RE-RANK, 8000",
|
|
|
|
| 2467 |
"tags": "RE-RANK, 4000",
|
| 2468 |
"max_tokens": 4000,
|
| 2469 |
"model_type": "rerank"
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"llm_name": "rerank-2",
|
| 2473 |
+
"tags": "RE-RANK, 16000",
|
| 2474 |
+
"max_tokens": 16000,
|
| 2475 |
+
"model_type": "rerank"
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"llm_name": "rerank-2-lite",
|
| 2479 |
+
"tags": "RE-RANK, 8000",
|
| 2480 |
+
"max_tokens": 8000,
|
| 2481 |
+
"model_type": "rerank"
|
| 2482 |
}
|
| 2483 |
]
|
| 2484 |
},
|
rag/llm/embedding_model.py
CHANGED
|
@@ -322,7 +322,7 @@ class YoudaoEmbed(Base):
|
|
| 322 |
|
| 323 |
|
| 324 |
class JinaEmbed(Base):
|
| 325 |
-
def __init__(self, key, model_name="jina-embeddings-
|
| 326 |
base_url="https://api.jina.ai/v1/embeddings"):
|
| 327 |
|
| 328 |
self.base_url = "https://api.jina.ai/v1/embeddings"
|
|
|
|
| 322 |
|
| 323 |
|
| 324 |
class JinaEmbed(Base):
|
| 325 |
+
def __init__(self, key, model_name="jina-embeddings-v3",
|
| 326 |
base_url="https://api.jina.ai/v1/embeddings"):
|
| 327 |
|
| 328 |
self.base_url = "https://api.jina.ai/v1/embeddings"
|
rag/llm/rerank_model.py
CHANGED
|
@@ -92,7 +92,7 @@ class DefaultRerank(Base):
|
|
| 92 |
|
| 93 |
|
| 94 |
class JinaRerank(Base):
|
| 95 |
-
def __init__(self, key, model_name="jina-reranker-
|
| 96 |
base_url="https://api.jina.ai/v1/rerank"):
|
| 97 |
self.base_url = "https://api.jina.ai/v1/rerank"
|
| 98 |
self.headers = {
|
|
|
|
| 92 |
|
| 93 |
|
| 94 |
class JinaRerank(Base):
|
| 95 |
+
def __init__(self, key, model_name="jina-reranker-v2-base-multilingual",
|
| 96 |
base_url="https://api.jina.ai/v1/rerank"):
|
| 97 |
self.base_url = "https://api.jina.ai/v1/rerank"
|
| 98 |
self.headers = {
|