Embedding configuration error

CrewAI Community Support
,
1

Hello guys,

I’m having a problem with embedding configuration.

here is my case :
I’m using PDFSearchTool and Memory.

My provider is Azure.

i saw that the embedder config for PDFSearchTool was not the same that RAGStorage (used for memory). The provider for PDFSearcheTool is “azure_openai” and for RAGStorage is “azure”.

So I need to do 2 differents config where only the provider changes.

The issue is that My config with RAGStorage is working perfectly, but with PDFSearchTool is not. So I have to use Ollama local model for PDFSearchTool and my Azure deployed model for RAGStorage.

I don’t have any error message execpt the response of agent saying “I am unable to complete the task due to technical issues with the PDF search tool. Please ensure the tool is properly configured and functional, or provide an alternative method for extracting information from the PDF.” stuffs like that.

Also the insert into ChromaDB Is not done :
Inserting batches in chromadb: 0%| | 0/1 [00:00<?, ?it/s]

that’s all I have.

Here is my configs :

EMBEDDER_CONFIG = dict(
provider="azure",
config=dict(
    model="text-embedding-ada-002",
    api_key=os.getenv("AZURE_API_KEY"),
    api_base=os.getenv("AZURE_API_BASE"),
    api_version=os.getenv("AZURE_API_VERSION"),
),

)

EMBEDDER_CONFIG_PDF = dict(
provider="azure_openai",
config=dict(
    model="text-embedding-ada-002",
    api_key=os.getenv("AZURE_API_KEY"),
    api_base=os.getenv("AZURE_API_BASE"),
),

)

For the Embedding it’s not AZURE_ON_PENAI_API_KEY but AZURE_API_KEY, in case you think it could be the cause.