The Developer's Guide to ChatGPT API
summary
Covers neural nets, Transformers and the attention idea, then moves to hands‑on use of the OpenAI API from a Google Colab notebook using Python, tiktoken and observablehq for token visualization. Demonstrations include embeddings search over 200 product reviews, DALL·E image generation, function calling that runs math correctly, moderation and a fine‑tune pipeline that produces an ft:gpt‑3.5 model and claims fine‑tune with ten examples is possible. Costs shown: GPT‑4 ≈ $10/1M tokens, GPT‑3.5 ≈ $0.50/1M; embeddings ≈ $0.02/1M.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 0:23Neural networks basics
- 7:51Text data prep & tokenization
- 16:20RNN limits for sequences
- 19:53Transformers & attention
- 27:39Encoder-decoder in GPT
- 30:52OpenAI API overview
- 35:50Embeddings model and its applications
- 36:12Fine-tuning models
- 42:32Q&A
- 53:17Hands-on Colab
- 1:00:07Function calling
- 1:03:18Embeddings for search
- 1:10:30Fine-tuning demo
- 1:22:57Moderation & final Q&A
related talks
01h 01mHow ChatGPT Works
01h 01mKeynote: Transformers: The Rise of ChatGPT
39mBuilding Your Own Bot with GPT: Harnessing the Power of AI...
01h 00mGoing beyond ChatGPT: introduction to prompt engineering & LLMs
28mGPT or how I learned to stop worrying and love the generative AI
46mUsing AI in Software Design: How ChatGPT Can Help With Creating a Solution Arch.
01h 09mChatGPT and OpenAI API solutions: 👑 My Favourite Implementations
16mUnveiling the practical prowess of OpenAI's ChatGPT