Pull Requests huggingface/peft

FIX Bug in forgetting metric in MetaMathQA

#3419 opened 2026-07-10 15:48 by BenjaminBossan

Improve IA3 documentation: add usage guidance, module selection tips, and LoRA comparison

#3417 opened 2026-07-10 01:02 by DresdenGman

ENH: Warn in get_peft_model when base model parameters are on the meta device

#3415 opened 2026-07-09 21:52 by immu4989

FIX DEFT's cached merge factor: precision loss and device staleness

#3412 opened 2026-07-09 06:21 by DaoyuanLi2816

FIX Pin huggingface_hub to <1.22

#3407 opened 2026-07-08 11:30 by BenjaminBossan

Implements EWoRA in PEFT

#3401 opened 2026-07-07 20:38 by HarshKohli

Fix broken link in GLoRA docs

#3397 opened 2026-07-07 00:37 by DaoyuanLi2816

Method comparison: Reduce peak memory usage of image generation task

#3395 opened 2026-07-06 15:05 by BenjaminBossan

Release 0.20.0

#3388 opened 2026-07-03 15:49 by BenjaminBossan

FEAT Add git dirty flag to method_comparison benchmark results

#3385 opened 2026-07-02 14:57 by peft-jambot

Add UniLoraConfig to ALL_CONFIG_CLASSES test matrix

#3383 opened 2026-07-01 21:39 by DaoyuanLi2816

Add Riemannian-preconditioned LoRA optimizer

#3382 opened 2026-07-01 19:00 by smellslikeml

FIX MiSS with bat init error with some quantization backends

#3381 opened 2026-07-01 16:23 by BenjaminBossan

fix: store miss_fn per-adapter in MiSS layer

#3379 opened 2026-07-01 12:46 by peft-jambot

chore: update torch_compile_tests.yml

#3378 opened 2026-07-01 10:56 by hf-security-analysis[bot]

CI: Bump the third-party-actions group across 1 directory with 7 updates dependencies github_actions

#3375 opened 2026-07-01 04:46 by dependabot[bot]

Add Relation Extraction LoRA script and notebook tutorial

#3368 opened 2026-06-28 10:08 by GauravPatil2515

feat: support LoRA conversion for multiplicative PEFT methods (OFT, BOFT)

#3360 opened 2026-06-25 09:47 by peft-jambot

Proposal: alternative ShadowPEFT implementation

#3357 opened 2026-06-24 15:57 by BenjaminBossan

refactor: move OFT to generic quantization framework

#3355 opened 2026-06-23 15:14 by peft-jambot

Feat: support ShadowPEFT

#3354 opened 2026-06-23 11:33 by SeanLee97

ENH Add PEFT method papers in a structured way

#3323 opened 2026-06-12 16:07 by BenjaminBossan

Minor app embed improvements

#3322 opened 2026-06-12 12:54 by githubnemo

Add quantization support for SHiRA

#3321 opened 2026-06-12 12:37 by githubnemo

fix: respect layer-specific LoRA-FA scaling in optimizer

#3315 opened 2026-06-10 22:05 by eSVeeF

fix(save_and_load): save base_layer.bias for bias="lora_only"

#3307 opened 2026-06-07 13:07 by Anai-Guo

Created tutorial detailing how to do differentially private PEFT fine-tuning of huggingface transformer models using FastDP

#3297 opened 2026-06-02 04:20 by lmarchyok

FIX: inject_adapter incorrectly propagates inference_mode to active adapters

#3290 opened 2026-05-30 07:58 by kiritozc

FIX BdLoraConfig.nblocks default value

#3252 opened 2026-05-22 02:48 by ATOM00blue

Fail loud on save_pretrained() for unsharded LoRA tensors

#3251 opened 2026-05-21 17:00 by akshansh47