Gain full control of your apps with Open-Source AI
summary
Claim: smaller open-source models often outcompete larger proprietary ones when distilled, fine-tuned and served with modern tooling. Evidence: Zier 7B hits top benchmarks despite 7B size; Hugging Face hosts ~140k text‑generation models and >1M total models; community labeling projects add 10k–22k examples with hundreds of contributors. Tools and results: Transformers + PEFT + quant libraries like bitsandbytes and GPTQ/AWQ plus text-generation-inference enable quantized, merged, edge-deployable models that actually run and cut inference cost.
AI-written summary · about
chapters (10)
- 0:00Open source AI importance
- 3:11Data quality vs model size
- 8:13Open source AI models
- 11:21Open-source progress examples
- 13:52Control and privacy with AI
- 16:21AI model deployment tools
- 19:56Fine-tuning with Transformers
- 22:10Parameter-efficient fine-tuning
- 23:34Quantization and optimization
- 24:38Let's build together
related talks
32mThe Rise of AI agents and open sources Ai models
20mTrue Exponential: Building a Generative AI Company
14mKeynote: Building AI with Open Source & Hugging Face- 14mKeynote: Building AI with Open Source & Hugging Face
52mBeyond the AI Hype: What's Real, What's Next
35mOpen-source AI: why it matters and how to get started
34mMapping the Open Model Landscape in 2025
27mKeynote: Open Source in the Age of AI