How to Develop APIs for Generative AI Drug Discovery Production

Here, 3 Large Language Model to Application Programming Interface methods are detailed. OPTION 1 features a Conda environment, an OpenLLM OpenAI-compatible API endpoint, and TII's Open Source Falcon model to deploy an API using terminal and Python scripts. OPTION 2 uses OpenLLM's getting started Jupyter Notebook and Google Colab to serve a Llama model. OPTION 3 utilizes the new LangServe Chat Playground to yield an API accessible by link - which is based on a LangChain template and GitHub repo creation. OPTIONs 1 and 3 are accompanied by seminars, and all three options feature the use of NVIDIA GPUs.
Event Seminar and PDF 23May24.

References:

04) Matthew Berman. (2023, June 22). API for Open-Source models Easily build with ANY Open-Source LLM [Video]. YouTube. https://www.youtube.com/watch?v=8nZZ2oQhx4E

05) Google Colab. (n.d.). https://colab.research.google.com/github/bentoml/OpenLLM/blob/main/examples/llama2.ipynb

09) LangChain. (2024c, March 12). LangServe Chat Playground Launch [Video]. YouTube. https://www.youtube.com/watch?v=bGUO26Bc-Sc

“We just need to know what we've been allowed to change or see, interfaces abstract away the implementation. Now, while the UI or user interface is made for the user of the application, the API is made for the application programmer to use and extend in their applications.”
Craig Dennis, APIs for Beginners 2023 - How to use an API (Full Course / Tutorial)