[Question]: Further explanation of total embedding token usage

[ihgumilar]

Question Validation

• [X] I have searched both the documentation and discord for an answer.

Question

Hi,

I have used this code https://docs.llamaindex.ai/en/stable/examples/callbacks/TokenCountingHandler/ to count for embedding token usage. However, I would like to get a further clarification or explanation do you have any specific breakdown about total embedding token usage in LlamaIndex ?

Does it mean the usage token for query and answer ? Further explanation is much appreciated about embedding token usage in LlamaIndex.

Thanks

[logan-markewich]

I'm not sure what you are looking for?

Indexing uses tokens to embed all your documents.

Querying uses much less, as it only needs to embed the query text

[ihgumilar]

Thanks for the explanation. Basically, I am trying to understand this

What does 8 (embedding tokens) represent ?

[logan-markewich]

It means to run the query, 8 tokens were sent to the embedding model (I.e it sent the query text to the embedding model)