Keynote: Building with Mistral
summary
Small Paris team launches the open-weight Mistral 7B (Sep 2023), follow-up mixture-of-experts models (8x7B and 22B) and commercial MrSmall/MrLarge offerings, plus a specialized 22B coding model trained on 80+ languages that beats larger models on benchmarks. Platform exposes an API, chat.m.ai interface, embeddings and a just-released fine-tuning service that adapts only small adapter matrices for cost and latency savings. Open-source repos include mr-inference, tokenizers, and a community cookbook with step-by-step fine-tuning and serving guides.
AI-written summary · about
chapters (6)
- 0:00Mistral AI introduction
- 1:08Mistral models overview
- 4:14Model customization services
- 6:22Codestral coding model
- 9:59Open-source tools
- 11:48Community engagement
related talks
40mBuilding with Mistral
34mA Journey into Open Source Models with HuggingFace & Mistral
29mLeveraging open-source LLMs for production
14mKeynote: Building AI with Open Source & Hugging Face
24mTeaching Mistral to Reason: Post-Training with PyTorch and NVIDIA
44mKeynote Speaker - Simon Willison
37mOpen-Source Generative AI: A Product Manager’s Blueprint
18mSmolLM: The Rise of Smol Models
