Keynote: Building AI with Open Source & Hugging Face
summary
Hugging Face positions itself as a mission-driven open platform and claims six surprising facts: over a million repositories with ~400,000 public models and ~10 million Transformer downloads per day; tooling includes Transformers, Optimum, PEFT and Accelerate; there are ~3,000 LLMs tracked and a searchable Open LLM Leaderboard. Production options include Text Generation Inference and inference endpoints (example: Mistral-like mixer available ~ $13/hr), Spaces for demos, HuggingChat as an open repo, browser JS libs and a Rust framework called Candle.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:41Hugging Face surprises
- 3:59Build AI with open source
- 7:51Build AI with Hugging Face
- 12:04Making AI accessible
- 13:29What Hugging Face is
- 14:11Closing
related talks
14mKeynote: Building AI with Open Source & Hugging Face
16mKeynote: Open Source AI Everywhere
40mKeynote: Open-source Multimodal AI in the Wild
01h 09mReinventing Machine Learning with Transformers and Hugging Face
11mKeynote: Building an Interoperable Ecosystem for Generative AI
13mKeynote: To Nemotron and Beyond: Building the Open Source AI Ecosystem
33mKeynote: Fine-tuning our way towards openness in AI
31mHugging Face ecosystem for Local AI/ ML