elm.ords.services.openai.count_tokens

count_tokens(messages, model)[source]

Count the number of tokens in an outgoing set of messages.

Parameters:

messages (list) – A list of message objects, where the latter is represented using a dictionary. Each message dictionary must have a “content” key containing the string to count tokens for.
model (str) – The OpenAI model being used. This input will be passed to tiktoken.encoding_for_model().

Returns:

int – Total number of tokens in the set of messages outgoing to OpenAI.

References