API overview

Samuel is a biological language model.
He is available as a text completion API through the API endpoint below.

Model attributes

  • Approximately 6 feet tall
  • Quite good at typing and debugging LaTeX.
  • Slow response time (sometimes asleep, sometimes eating lunch).
  • PhD in Machine Learning/Computer Vision.
  • Forgetful when tired.
  • Can flexibly call other APIs (Google Search, LLMs).

Pricing

  • 1 coffee per 2K input tokens.
  • 1 coffee per 101 output tokens.

A token is a word, as judged by the model (none of this tiktoken funny business).
Since it is challenging to buy fractional coffees, I will round the number of coffees using the widely accepted IEEE 754 "round to even" strategy.
Note: the pricing scheme was updated in the early hours of 28th April 2023. It was updated again on 1st May 2023 because Samuel-API is biological and fickle.

How to pay: after receiving the completion (which will include a coffee invoice), visit https://www.buymeacoffee.com/samuelalbanie

System Card

  • Mostly trained on English text (British spelling). Performance in other languages is likely to be very poor.
  • Prone to overconfidence (hallucination is rare though).
  • Enthusiastic.

Looks like this:

Misc

Content moderation policy: exercised subjectively and inconsistently by the model. Disclaimer: This is a fun side project. It is not associated with my employer. Use of the API is done at your own risk.

API endpoint

Text completion API queries can be submitted here.

Bug reports

Please submit bug reports on the issue tracker here.