From acceleration knobs and extreme quantization to a nine-model comparison, a large-sample retest at month's end reshuffled the entire leaderboard. A record of a broken yardstick, small-sample error bars, and the current division of labor between pi and local models.
Run LoRA local fine-tuning of Gemma-4 E4B on an M4 Pro with mlx-tune. In 33 seconds the model learns your JSON schema. A full field report of Apple Silicon local fine-tuning.
Fine-tune Qwen3.5 with QLoRA on an RTX 4060 Ti using 76 ChatML samples. A full walkthrough from environment setup through LoRA merge, GGUF conversion, and Ollama deploy, plus five hard rules for Windows CUDA fine-tuning.
A hands-on guide to deploying a LINE Bot customer service agent on Windows GPU with Ollama, Qwen3.5, and Hermes Gateway, including six real production pitfalls and fixes
LoRA and QLoRA let you fine-tune a large language model with just a few dozen examples on your own computer, no cloud GPU rental, no uploading sensitive data. A real biotech lab case shows why this technique changes how local AI feels to use.
I benchmarked five 4-bit local models on a single M4 Pro (48GB) with oMLX across eight benchmarks, mapped the speed/quality sweet spot, and calibrated which cloud generation they match. Plus a field report on three speculative-decoding traps (MTP / dflash / specprefill) and the cache trade-off that decides them.