huggingface/peft issues and pull requests

#1480 - FIX [`CI` / `bnb`] Fix failing bnb workflow

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 4 comments

#1479 - Initial Commit to implement Dora Adapter based on LoRA

Pull Request - State: closed - Opened by napsternxg 8 months ago - 2 comments

#1478 - Connection to huggingface.co timed out when merging Qwen model

Issue - State: closed - Opened by anyiz 8 months ago - 3 comments

#1477 - Allow trust_remote_code for tokenizers when loading AutoPeftModels

Pull Request - State: closed - Opened by OfficialDelta 8 months ago - 3 comments

#1476 - AQLM support for LoRA

Pull Request - State: closed - Opened by BlackSamorez 8 months ago - 11 comments

#1475 - Update docstring at peft_types.py

Pull Request - State: closed - Opened by eduardozamudio 8 months ago - 1 comment

#1474 - Implement DoRA

Pull Request - State: closed - Opened by BenjaminBossan 8 months ago - 50 comments

#1472 - Integrating X-LoRA

Issue - State: closed - Opened by EricLBuehler 8 months ago - 15 comments

#1471 - Adds an example notebook for showing multi-adapter weighted inference

Pull Request - State: closed - Opened by sayakpaul 8 months ago - 2 comments

#1470 - [BUG] Unexpected GPU memory consumption when using transformers PEFT in Zero3

Issue - State: closed - Opened by alekseymalakhov11 8 months ago - 3 comments

#1469 - Using LoRA consumes high memory

Issue - State: closed - Opened by WenxiongLiao 8 months ago - 12 comments

#1468 - Using Pipeline and TGI having the LoRA Adapter.

Issue - State: closed - Opened by wolfassi123 8 months ago - 3 comments

#1467 - Add pre-commit configuration

Pull Request - State: closed - Opened by akx 8 months ago - 2 comments

#1466 - add `magnitude_prune` merging method

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1465 - [`CI`] Fix adaptation prompt CI on transformers main

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 2 comments

#1463 - Make tests succeed more on MPS

Pull Request - State: closed - Opened by akx 8 months ago - 4 comments

#1462 - Ability to add new tokens and keep existing token embeddings frozen?

Issue - State: closed - Opened by DreamGenX 8 months ago - 9 comments

#1461 - [`CI`] Add CI tests on transformers main to catch early bugs

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 2 comments

#1460 - The error occurs when there is no argument --with_tracking from peft_lora_embedding_semantic_search code

Issue - State: closed - Opened by smileinthebluesky 8 months ago - 2 comments

#1459 - [ `Adaptation Prompt`] Fix llama rotary embedding issue with transformers main

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 1 comment

#1458 - [`core` / `get_peft_state_dict`] Ignore all exceptions to avoid unexpected errors

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 3 comments

#1457 - Size mismatch error when loading model with added tokens using peft_model.save_pretrained

Issue - State: closed - Opened by Rizvanov-Rinat 8 months ago - 4 comments

#1456 - [docs] Remove iframe

Pull Request - State: closed - Opened by stevhliu 8 months ago - 1 comment

#1455 - [docs] Docstring typo

Pull Request - State: closed - Opened by stevhliu 8 months ago - 1 comment

#1454 - FIX Honor HF_HUB_OFFLINE mode if set by user

Pull Request - State: closed - Opened by BenjaminBossan 8 months ago - 7 comments

#1453 - RunTimeError Missing keys while resuming training and cannot load checkpoint

Issue - State: closed - Opened by seanbenhur 8 months ago - 14 comments

#1452 - peft/utils/save_and_load.py try to connect to the hub even when HF_HUB_OFFLINE=1

Issue - State: closed - Opened by LsTam91 8 months ago - 1 comment

#1451 - FIX: fix for dimension mismatch when using adaption prompt in Llama2

Pull Request - State: closed - Opened by LEE94-1 8 months ago - 5 comments

#1450 - Support `modules_to_save` config option when using DeepSpeed ZeRO-3 with ZeRO init enabled.

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1449 - FIX Loading with AutoPeftModel.from_pretrained

Pull Request - State: closed - Opened by BenjaminBossan 8 months ago - 1 comment

#1448 - Use plain asserts in tests

Pull Request - State: closed - Opened by akx 8 months ago - 5 comments

#1447 - Add docstrings for set_adapter and keep frozen

Pull Request - State: closed - Opened by EricLBuehler 8 months ago - 11 comments

#1446 - MNT: Check only selected directories with ruff

Pull Request - State: closed - Opened by BenjaminBossan 8 months ago - 2 comments

#1445 - TST: Improve test coverage by skipping fewer tests

Pull Request - State: closed - Opened by BenjaminBossan 8 months ago - 1 comment

#1444 - [Docs] call `set_adapters()` after add_weighted_adapter

Pull Request - State: closed - Opened by sayakpaul 8 months ago - 1 comment

#1443 - size mismatch for base_model.model.model.layers.0.mlp.gate_proj.lora_B.default.weight: copying a param with shape torch.Size([0]) from checkpoint, the shape in current model is torch.Size([14336, 16]).

Issue - State: closed - Opened by tamanna-mostafa 8 months ago - 16 comments

#1442 - RuntimeError when combining FSDP and disable_adapter

Issue - State: closed - Opened by wanghao14 8 months ago - 3 comments

#1440 - Fixed saving for models that don't have _name_or_path in config

Pull Request - State: closed - Opened by kovalexal 8 months ago - 1 comment

#1439 - [`core`/`TPLinear`] Fix breaking change

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 2 comments

#1438 - ValueError: Tokenizer class XXXXXXXX does not exist or is not currently imported

Issue - State: closed - Opened by Sun-Shiqi 8 months ago - 6 comments

#1437 - Update Dockerfile to reflect how to compile bnb from source

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 1 comment

#1436 - [WIP] Adding support for Embedding4bit in bnb.py at tuners/lora

Pull Request - State: closed - Opened by GM-git-dotcom 8 months ago - 21 comments

#1435 - Fix typos

Pull Request - State: closed - Opened by szepeviktor 8 months ago - 3 comments

#1434 - Running multiple adapters simultaneously, reusing the output from the base model

Issue - State: closed - Opened by sachin-patel-qp 8 months ago - 4 comments

#1433 - Add Support for Mistral Model in Llama-Adapter Method

Pull Request - State: closed - Opened by PrakharSaxena24 8 months ago - 9 comments

#1432 - ValueError: Cannot flatten integer dtype tensors

Issue - State: closed - Opened by fancyerii 8 months ago - 14 comments

#1431 - Extend PeftModel.from_pretrained() to models with disk-offloaded modules

Pull Request - State: closed - Opened by blbadger 8 months ago - 14 comments

#1430 - huggingface_hub.utils._validators.HFValidationError when AutoPeftModelForCausalLM.from_pretrained

Issue - State: closed - Opened by fancyerii 8 months ago - 2 comments

#1428 - Release patch version 0.8.2

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1427 - [`core`] fix critical bug in diffusers

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 1 comment

#1426 - Significant difference in size in lora artifacts between different versions of peft

Issue - State: closed - Opened by debraj135 8 months ago - 1 comment

#1425 - Fix `MatMul8bitLtBackward` view issue

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 6 comments

#1424 - PeftModel error with forward()

Issue - State: closed - Opened by lqf0624 8 months ago - 8 comments

#1423 - [docs] Model merging

Pull Request - State: closed - Opened by stevhliu 8 months ago - 3 comments

#1422 - [docs] Polytropon API

Pull Request - State: closed - Opened by stevhliu 8 months ago - 1 comment

#1421 - Renovate quality tools

Pull Request - State: closed - Opened by akx 8 months ago - 3 comments

#1420 - DO NOT MERFGE: Attempt to fix memory

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 4 comments

#1419 - How to torch.jit.trace a peft model

Issue - State: closed - Opened by dcy0577 8 months ago - 3 comments

#1418 - [Documentation] Compatibility between lora + deepspeed + bitsandbytes

Issue - State: closed - Opened by nathan-az 8 months ago - 9 comments

#1417 - [Hub] fix: subfolder existence check

Pull Request - State: closed - Opened by sayakpaul 8 months ago - 1 comment

#1416 - Release v0.8.2.dev0

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1415 - Patch Release v0.8.1

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1414 - Fix breaking change

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 6 comments

#1413 - Fix PEFT breaking change

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 1 comment

#1412 - Release 0.8.1.dev0

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1411 - [docs] README update

Pull Request - State: closed - Opened by stevhliu 8 months ago - 2 comments

#1410 - no trainable params when set adapter_name

Issue - State: closed - Opened by dkqkxx 8 months ago - 1 comment

#1409 - Update custom_models.md

Pull Request - State: closed - Opened by boyufan 8 months ago - 1 comment

#1408 - Fix various typos in LoftQ docs.

Pull Request - State: closed - Opened by arnavgarg1 8 months ago - 2 comments

#1407 - Add IA3 Modules for Phi

Pull Request - State: closed - Opened by arnavgarg1 8 months ago - 2 comments

#1406 - Release: v0.8.0

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1405 - loftq example non-functional - assert absmax is not None and out is not None

Issue - State: closed - Opened by winglian 8 months ago - 6 comments

#1404 - [Documentation] Request to Include LoftQ and LQ-LoRA

Issue - State: closed - Opened by HanGuo97 8 months ago - 5 comments

#1403 - Simplify `PeftModel.save_pretrained`

Issue - State: closed - Opened by simon-lund 8 months ago - 4 comments

#1402 - LoRA adapters weights remain zero after training with deepspeed

Issue - State: closed - Opened by schwarzwalder93 8 months ago - 7 comments

#1401 - Where is `self.generation_config`coming from?

Issue - State: closed - Opened by simon-lund 8 months ago - 7 comments

#1400 - New pypi version please

Issue - State: closed - Opened by YLGH 8 months ago - 1 comment

#1399 - FEAT: add awq suppot in PEFT

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 2 comments

#1398 - Add generic PeftConfig constructor from kwargs

Pull Request - State: closed - Opened by sfriedowitz 8 months ago - 2 comments

#1397 - Why huggingface only support relative local path?

Issue - State: closed - Opened by yzeng58 8 months ago - 4 comments

#1396 - RuntimeError: Error(s) in loading state_dict for PeftModelForCausalLM: size mismatch...

Issue - State: closed - Opened by achangtv 8 months ago - 5 comments

#1395 - Missing Keys when loading checkpoints using deepspeed and HuggingFace Trainer

Issue - State: closed - Opened by nathan-az 8 months ago - 1 comment

#1394 - [docs] Doc maintenance

Pull Request - State: closed - Opened by stevhliu 8 months ago - 2 comments

#1393 - Add positional args to PeftModelForCausalLM.generate

Pull Request - State: closed - Opened by SumanthRH 8 months ago - 1 comment

#1392 - [Q] Does support for embedding and conv2d layers in bnb.py at tuners/lora make sense?

Issue - State: closed - Opened by GM-git-dotcom 8 months ago - 8 comments

#1391 - Softprompt broken with use_cache=True

Issue - State: closed - Opened by Rallio67 8 months ago - 3 comments

#1390 - Clm Prompt tuning inference based on tutorial throwing runtime error

Issue - State: closed - Opened by LaviRoars 8 months ago - 4 comments

#1389 - models are the same after loading lora parameters using peft library

Issue - State: closed - Opened by ra-MANUJ-an 8 months ago - 5 comments

#1388 - merge_and_unload() raises the loss of accuracy when inference

Issue - State: closed - Opened by haobozhang 8 months ago - 2 comments

#1386 - [docs] Minor fixes to tutorial on semantic segmentation with LoRA

Pull Request - State: closed - Opened by GM-git-dotcom 8 months ago - 2 comments

#1385 - [docs] Add CONTRIBUTING.md to repo

Pull Request - State: closed - Opened by aaronjimv 8 months ago - 3 comments

#1384 - Saving and Loading fine-tuned model

Issue - State: closed - Opened by binbinxue 8 months ago - 3 comments

#1383 - save the embeddings even when they aren't targetted but resized

Pull Request - State: closed - Opened by pacman100 8 months ago - 1 comment

#1382 - How to set a predefined weight for LoRA and the linear layer

Issue - State: closed - Opened by quqxui 8 months ago - 2 comments

#1381 - wrong model loading after merge_and_unload + save_pretrained

Issue - State: closed - Opened by sd3ntato 8 months ago - 5 comments

#1380 - add mixtral in LoRA mapping

Pull Request - State: closed - Opened by younesbelkada 8 months ago - 1 comment

#1379 - Not able to load peft (promt-tuned) model in multi-gpu settings for inference

Issue - State: closed - Opened by dineshkh 8 months ago - 6 comments

GitHub / huggingface/peft issues and pull requests