Fine-tuning an LLM to write docs like it's 1995
Hacker News (AI keywords)9 days agoTutorialThe author builds a corpus from old Microsoft manuals, cleans OCR text, generates instruction-style JSONL examples, and fine-tunes Llama 3.1 8B and Qwen 2.5 7B with QLoRA. Tests cover malloc(), a fictional Win32 API, and a deliberately anachronistic REST API prompt. Qwen fine-tunes transfer the period documentation style best, but the experiment also shows hallucination risks, tuning complexity, and why these models augment rather than replace technical writers.
在消費級硬體上微調 FLUX.1-dev：使用 QLoRA 技術指南★ 80
Hugging Face Blog360 days agoTutorialFLUX.1-dev is a state-of-the-art open-source text-to-image model with 12 billion parameters (12B), developed by Black Forest Labs. However, due to its enormous…
在 Hugging Face 中微調 Gemma 模型★ 80
Hugging Face Blog842 days agoTutorialAfter Google released the Gemma family of open-source models (including 2B and 7B parameter versions), Hugging Face promptly published this practical…
使用 Unsloth 與 🤗 TRL 讓 LLM 微調速度提升 2 倍★ 80
Hugging Face Blog886 days agoReleaseHugging Face's official blog announced a partnership with the Unsloth team to integrate Unsloth's efficient fine-tuning technology directly into Hugging Face's…
2023 年：開源大語言模型（Open LLMs）爆發之年★ 75
Hugging Face Blog909 days agoCommentaryLooking back on 2023, the most notable trend in the AI landscape was the explosive growth of open-source large language models (Open LLMs). In this annual…
Hugging Face Transformers 原生支援量化方案全解析：bitsandbytes 與 GPTQ 實戰指南★ 75
Hugging Face Blog1,006 days agoTutorialAs the parameter count of large language models (LLMs) has grown dramatically, running and fine-tuning these models on consumer-grade GPUs or limited hardware…
Hugging Face 整合 bitsandbytes、4-bit 量化與 QLoRA，讓大型語言模型更親民★ 90
Hugging Face Blog1,117 days agoReleaseThis official Hugging Face blog post introduces a deep integration with the `bitsandbytes` library, formally adding 4-bit quantization support to…