Fine-tuning
Pre-training, post-training, fine-tuning, distillation, quantization and the data that feeds them
Latest stories See all →
- RAG vs Fine-Tuning: Which One Do You Actually Need?
- I’m not an engineer. I fine-tuned my own language model on a MacBook Air
- I Reproduced Pinterest’s Quantization Numbers on a Laptop, and Accidentally Worked Out Their…
- Bolt is giving developers 50x more compute. But there’s a catch.
- 5 Free Microsoft GitHub Courses to Learn Data Science and Artificial Intelligence
- Same bytes, closer to the original: two lines of AutoRound we had wrong
- When Should You Fine-Tune Instead of Prompt-Engineer?
- What Is an LLM Model Registry and How Is It Different from an ML Model Registry?
- What If the Adaptation Were a Model? ShadowPEFT in 🤗 PEFT library
- Fine-Tuning vs Prompting: When to Specialize an SLM
- Run GLM 5.3 Flash Locally: GSQ and RCO Quantization Explained
- Accelerating Sovereign AI: Domyn's Journey with NVIDIA
- Foundation Models in Radiology: What Data Do You Actually Need to Fine-Tune Them?
- A Practical LLM Pretraining Pipeline with LanceDB
- Understanding W8A8 INT8 LLM quantization: Accuracy and performance results
- Run 40% more post-training experiments on the same GPUs with llm-d time-slicing
- Same bytes, closer to the original: two lines of AutoRound we had wrong
- Looking for lateral movement with a neural network trained on synthetic data
- Is building a manual evaluation set the right approach when no reliable labeled ground truth exists for resume-job matching?
- NeoHorse-1-4B: How to Run This Self-Improving 4B Model Locally
- Recursive Self-Improvement Training: How NeoHorse-1-4B Learns From Itself
- Rubric Design: The Missing Layer Between Guidelines and Good Annotations
- Edge0-35B Benchmarks: What 4-bit Quantization Really Costs You
- How to Generate Labeled UI Interaction Data at Scale With Synthetic Data
Podcasts
Recent episodes
- AI Safety Alarms, China's Distillation Reckoning, and Qualcomm's AWS Breakthrough: A Pivotal Week for Enterprise AI
- Fusion power startups find new partners in the defense world; plus, Mecka AI nears $500M valuation amid rush for robot training data
- It's A Doozy
- Risky Bulletin: Ukraine's top prosecutor resigns amid scam call center scandal
- Risky Business #852 -- Cyber Command wants to buy shells
- Java’s age is its AI superpower
- Aaron Levie on Why Open AI Wins
- Aaron Levie on Why Open AI Wins
- Aaron Levie on Why Open AI Wins
- Nvidia Buys Hugging Face For A Rabbit?
- #255 - Gemini 3.7, Jalapeño, Qwen 3.8, Drones
- SN 1093: Tokens in the Stream - Why LLMs are inherently insecure and prompt injection will persist