Krasus.Chen

yes ,on azure just embed one by one , ``` ["id1"],["meta1"],["doc1"] ``` wrong example : ``` ["id1",""id2"],["meta1","meta2"],["doc1","doc2"] ``` you will get : Too many inputs. The max number of inputs...

tks， I upgrade langchain ,and it's gone..

合并是肯定的，gpt应该会结合上下文回复最后一个问题。

show your codes and logs , please.

> looks like cassandra chat history has recently been updated, not sure if this pr still applies @lingfengchencn sure, it's for old version.

just set 'engine' or 'deployment_id' ``` #deployment_id or engine you can get from azure AzureOpenai(deployment_name="gpt-3.5-turbo") ```

#5345 same issue try code: ```python openai.api_type = "azure" openai.api_key = API_KEY openai.api_base = RESOURCE_ENDPOINT openai.api_version = '2023-03-15-preview' url = openai.api_base + f"/openai/deployments?api-version={openai.api_version }" r = requests.get(url, headers={"api-key": API_KEY}) print(r.text)...

show your deployments result,plz. make sure your deployment_name is "text-davanci".

ids are your deployment_name "model": "text-davinci-003", "owner": "organization-owner", **"id": "Otext-davanci",** "model": "gpt-35-turbo", "owner": "organization-owner", **"id": "gpt-35-turbo",** "model": "text-embedding-ada-002", "owner": "organization-owner", **"id": "text-embedding-ada-002",** you can try : ```python llm = AzureOpenAI(...

Sorry, I have no idea. can you use openai's sdk to test whether the engine / deployment_id / deployment_name(in langchain) is valid? ```python import openai openai.api_key = '...' openai.api_base =...