{"slug": "peft", "name": "PEFT", "category": "dev-libs", "category_name": "Libraries and frameworks", "description": "", "recommenders": 3, "commenters": 3, "warners": 0, "score": 2.695, "rising": -0.98, "first": "2026-04-01", "latest": "2026-06-14", "why": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 3, "neutral": 0, "compare": 0, "warn": 0}, "spark": [0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0], "listed": true, "loved": [], "caveat": "", "short_quote": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs", "kind": "", "icon": false, "domain": null, "github_repo": "github.com/huggingface/peft", "official_url": "https://huggingface.co/docs/peft", "pricing": "", "aliases": ["PEFT"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 776489216, "item_id": 48530576, "url": "https://news.ycombinator.com/item?id=48530576", "by": "calebkaiser", "date": "2026-06-14", "version": null, "stance": "recommend", "quote": "This is a good starting point", "excerpt": "This is a good starting point: https://huggingface.co/docs/peft/developer_guides/model_merging But yes, in general, merging refers to techniques that directly blend the weights of different models mathematically. It had a big moment of popularity ~2 years ago, with many so-called \"Frankenmodels\" popping up on leaderboards. I tend to think of merging as belonging to the same general umbrella as things like \"abliterati", "context": "", "use_case": "model merging docs", "switched_from": "", "story": "Rio de Janeiro's \"homegrown\" LLM appears to be a merge of an existing model"}, {"id": 772761360, "item_id": 48297585, "url": "https://news.ycombinator.com/item?id=48297585", "by": "minimaxir", "date": "2026-05-27", "version": null, "stance": "recommend", "quote": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs", "excerpt": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs, but for a machine learning engineer working with LLM models outside of just prompting an API, it's within scope to know why/how a LoRA should be used.", "context": "", "use_case": "creating LoRA adapters", "switched_from": "", "story": "The worst job interview I ever had"}, {"id": 761610448, "item_id": 47600653, "url": "https://news.ycombinator.com/item?id=47600653", "by": "kashifr", "date": "2026-04-01", "version": null, "stance": "recommend", "quote": "You can try out TinyLoRA in PEFT main now", "excerpt": "You can try out TinyLoRA in PEFT main now: https://huggingface.co/docs/peft/main/en/package_reference/tinylora", "context": "", "use_case": "", "switched_from": "", "story": "TinyLoRA \u2013 Learning to Reason in 13 Parameters"}]}], "caveats": [], "who": [], "tags": [], "use_cases": ["model merging docs", "creating lora adapters"], "compared": [], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=48528371", "title": "Rio de Janeiro's \"homegrown\" LLM appears to be a merge of an existing model", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48285344", "title": "The worst job interview I ever had", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=47541733", "title": "TinyLoRA \u2013 Learning to Reason in 13 Parameters", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {}, "2026-04": {"recommend": 1}, "2026-05": {"recommend": 1}, "2026-06": {"recommend": 1}, "2026-07": {}, "2026-08": {}, "2026-09": {}, "2026-10": {}}, "versions": [], "version_unnamed": 0, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 776489216, "item_id": 48530576, "url": "https://news.ycombinator.com/item?id=48530576", "by": "calebkaiser", "date": "2026-06-14", "version": null, "stance": "recommend", "quote": "This is a good starting point", "excerpt": "This is a good starting point: https://huggingface.co/docs/peft/developer_guides/model_merging But yes, in general, merging refers to techniques that directly blend the weights of different models mathematically. It had a big moment of popularity ~2 years ago, with many so-called \"Frankenmodels\" popping up on leaderboards. I tend to think of merging as belonging to the same general umbrella as things like \"abliterati", "context": "", "use_case": "model merging docs", "switched_from": "", "story": "Rio de Janeiro's \"homegrown\" LLM appears to be a merge of an existing model"}, {"id": 772761360, "item_id": 48297585, "url": "https://news.ycombinator.com/item?id=48297585", "by": "minimaxir", "date": "2026-05-27", "version": null, "stance": "recommend", "quote": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs", "excerpt": "PEFT is more niche as it's a specific library (https://github.com/huggingface/peft) that creates LoRAs, but for a machine learning engineer working with LLM models outside of just prompting an API, it's within scope to know why/how a LoRA should be used.", "context": "", "use_case": "creating LoRA adapters", "switched_from": "", "story": "The worst job interview I ever had"}, {"id": 761610448, "item_id": 47600653, "url": "https://news.ycombinator.com/item?id=47600653", "by": "kashifr", "date": "2026-04-01", "version": null, "stance": "recommend", "quote": "You can try out TinyLoRA in PEFT main now", "excerpt": "You can try out TinyLoRA in PEFT main now: https://huggingface.co/docs/peft/main/en/package_reference/tinylora", "context": "", "use_case": "", "switched_from": "", "story": "TinyLoRA \u2013 Learning to Reason in 13 Parameters"}]}