LookupError: Model(bge-reranker-v2-m3) not authorized

[github-eliviate]

问题： 如何获取rerank_id？

所做工作：

1. 浏览器访问v1/models输出： {"object":"list","data":[{"id":"bge-reranker-v2-m3","object":"model","created":1765166385,"owned_by":"vllm","root":"/model","parent":null,"max_model_len":8192,"permission":[{"id":"modelperm-45224c9ca3bd41e8bff883d8bdefdf47","object":"model_permission","created":1765166385,"allow_create_engine":false,"allow_sampling":true,"allow_logprobs":true,"allow_search_indices":false,"allow_view":true,"allow_fine_tuning":false,"organization":"*","group":null,"is_blocking":false}]}]}

调用的python api: rag_object.retrieve(question=item["search_query"],dataset_ids=dataset_ids,similarity_threshold=0.4,top_k=30,rerank_id="bge-reranker-v2-m3")

报错：

  File "/ragflow/api/apps/sdk/doc.py", line 1453, in retrieval_test
    rerank_mdl = LLMBundle(kb.tenant_id, LLMType.RERANK, llm_name=req["rerank_id"])
  File "/ragflow/api/db/services/llm_service.py", line 87, in __init__
    super().__init__(tenant_id, llm_type, llm_name, lang, **kwargs)
  File "/ragflow/api/db/services/tenant_llm_service.py", line 256, in __init__
    self.mdl = TenantLLMService.model_instance(tenant_id, llm_type, llm_name, lang=lang, **kwargs)
  File "/ragflow/.venv/lib/python3.10/site-packages/peewee.py", line 3128, in inner
    return fn(*args, **kwargs)
  File "/ragflow/api/db/services/tenant_llm_service.py", line 138, in model_instance
    model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
  File "/ragflow/.venv/lib/python3.10/site-packages/peewee.py", line 3128, in inner
    return fn(*args, **kwargs)
  File "/ragflow/api/db/services/tenant_llm_service.py", line 132, in get_model_config
    raise LookupError("Model({}) not authorized".format(mdlnm))
LookupError: Model(bge-reranker-v2-m3) not authorized

[dosubot[bot]]

Hi @github-eliviate! I'm Dosu and I’m helping the ragflow team.

你的报错是因为 rerank_id 没有被正确注册或授权给当前 tenant。rerank_id 必须和后台注册的模型名称完全一致，通常需要带上厂商后缀，比如 bge-reranker-v2-m3@VLLM 或 bge-reranker-v2-m3___VLLM，而不是直接用 bge-reranker-v2-m3。你可以在 UI 的“Retrieval Testing”页面的“Rerank model”下拉菜单里看到所有有效的 rerank_id，只有这些才是被授权的模型 ID【参考1】。

如果你用的是自定义或本地部署的模型，需要管理员在后台（比如通过 /add_llm 接口或管理界面）为当前 tenant 注册并启用 rerank 模型，指定正确的 llm_factory（如 VLLM）和 llm_name（如 bge-reranker-v2-m3___VLLM），并确保 status=1 才能被授权【参考2】。

常见问题包括：rerank_id 拼写不一致、缺少厂商后缀、tenant_llm 表里有冲突或失效的记录等。可以尝试用带后缀的 rerank_id，如果还是不行，建议让管理员检查 tenant_llm 表或重新注册模型。

参考：

• rerank_id 必须和注册名完全一致，通常带有厂商后缀
• 模型注册和授权机制说明
• rerank_id 获取和常见问题

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[Magicbook1108]

I think dosu is correct

[github-eliviate]

I think dosu is correct

你这个正确，昨天同事也是从这里帮我找到的，但是我不懂前端，目前还不知道是否能通过其它渠道得到这个rerank_id

[github-eliviate]

Retrieval Testing”页面的“Rerank model”下拉菜单里看到所有有效的 rerank_id

传菜单的不对，缺少后缀

厂商后缀 厂商也不对，我用vllm部署的，实际上rerank_id是 bge-reranker-v2-m3___OpenAI-API@OpenAI-API-Compatible

[github-eliviate]

建议应该有一个类似界面的东西，可以查询rerank_id，复制过来可以直接用的那种

[Magicbook1108]

Yeah, I strongly suggest you to raise a feature request to show dataset_id, document_id, chat_id, session_id, agent_id ... all these ids that we need in Http api and python sdk.