huggingface/peft issues and pull requests

#875 - Merge lora module to 8bit model

Pull Request - State: open - Opened by jiqing-feng about 1 year ago - 7 comments

#874 - Fix error using deepspeed zero2 + load_in_8bit + lora

Pull Request - State: closed - Opened by tmm1 about 1 year ago - 3 comments

#874 - Fix error using deepspeed zero2 + load_in_8bit + lora

Pull Request - State: closed - Opened by tmm1 about 1 year ago - 3 comments

#873 - support multiple ranks and alphas for LoRA

Pull Request - State: closed - Opened by pacman100 about 1 year ago - 11 comments

#872 - Faster init with LoRA

Pull Request - State: closed - Opened by poedator about 1 year ago - 18 comments

#872 - Faster init with LoRA

Pull Request - State: open - Opened by poedator about 1 year ago - 16 comments

#871 - unnecessarily slow get_peft_model() with LoRA (esp. in LLMs)

Issue - State: closed - Opened by poedator about 1 year ago - 11 comments
Labels: enhancement

#871 - unnecessarily slow get_peft_model() with LoRA (esp. in LLMs)

Issue - State: open - Opened by poedator about 1 year ago - 8 comments
Labels: enhancement

#870 - RuntimeError in Prefix

Issue - State: closed - Opened by liuao743 about 1 year ago - 8 comments

#870 - RuntimeError in Prefix

Issue - State: open - Opened by liuao743 about 1 year ago - 1 comment

#869 - How to correctly use Prefixing Tuning?

Issue - State: closed - Opened by Vincent-Li-9701 about 1 year ago - 33 comments

#868 - merge_and_unload issue?

Issue - State: closed - Opened by Remmet577 about 1 year ago - 18 comments
Labels: bug

#868 - merge_and_unload issue?

Issue - State: open - Opened by Remmet577 about 1 year ago

#867 - Ln437 in peft_model : hub_file_exists will cause unhandled exception when TRANSFORMERS_OFFLINE=1

Issue - State: open - Opened by gamert about 1 year ago - 1 comment

#867 - Ln437 in peft_model : hub_file_exists will cause unhandled exception when TRANSFORMERS_OFFLINE=1

Issue - State: closed - Opened by gamert about 1 year ago - 2 comments

#864 - Add 4-bit support to IA3 - Outperforms QLoRA in both speed and memory consumption

Pull Request - State: open - Opened by His-Wardship about 1 year ago - 4 comments

#864 - Add 4-bit support to IA3 - Outperforms QLoRA in both speed and memory consumption

Pull Request - State: closed - Opened by His-Wardship about 1 year ago - 23 comments

#862 - Inference with DDP will allocate multiplied lora weights to gpu rank 0

Issue - State: closed - Opened by Honesty-of-the-Cavernous-Tissue about 1 year ago - 2 comments

#862 - Inference with DDP will allocate multiplied lora weights to gpu rank 0

Issue - State: open - Opened by Honesty-of-the-Cavernous-Tissue about 1 year ago

#860 - ImportError: cannot import name 'is_npu_available' from 'accelerate.utils'

Issue - State: closed - Opened by mahbubur-87 about 1 year ago - 2 comments

#860 - ImportError: cannot import name 'is_npu_available' from 'accelerate.utils'

Issue - State: closed - Opened by mahbubur-87 about 1 year ago - 2 comments

#859 - Unable to load optimizer checkpoint when using resume_from_checkpoint

Issue - State: open - Opened by activatedgeek about 1 year ago - 6 comments

#859 - Unable to load optimizer checkpoint when using resume_from_checkpoint

Issue - State: closed - Opened by activatedgeek about 1 year ago - 6 comments

#858 - feat: add type hints

Pull Request - State: open - Opened by SauravMaheshkar about 1 year ago

#858 - feat: add type hints

Pull Request - State: closed - Opened by SauravMaheshkar about 1 year ago - 6 comments

#857 - How to load fine tune IDEFICS model with peft for inference?

Issue - State: closed - Opened by imrankh46 about 1 year ago - 15 comments

#857 - How to load fine tune IDEFICS model with peft for inference?

Issue - State: open - Opened by imrankh46 about 1 year ago - 2 comments

#856 - Keep getting "has no attribute '_get_submodules'" trying to apply model

Issue - State: closed - Opened by practical-dreamer about 1 year ago - 9 comments

#856 - Keep getting "has no attribute '_get_submodules'" trying to apply model

Issue - State: open - Opened by practical-dreamer about 1 year ago - 7 comments

#855 - Distributed training not working for PEFT/AdaLoRA whisper

Pull Request - State: open - Opened by rrva about 1 year ago - 3 comments

#855 - Distributed training not working for PEFT/AdaLoRA whisper

Pull Request - State: closed - Opened by rrva about 1 year ago - 8 comments

#854 - Sharing backbone model for different task adapters

Issue - State: closed - Opened by jedzill4 about 1 year ago - 2 comments

#854 - Sharing backbone model for different task adapters

Issue - State: open - Opened by jedzill4 about 1 year ago - 1 comment

#853 - DOC: Remove backlog section from README.md

Pull Request - State: closed - Opened by BenjaminBossan about 1 year ago - 1 comment

#852 - MergedLinear still available?

Issue - State: open - Opened by DanielRoeder1 about 1 year ago - 2 comments

#852 - MergedLinear still available?

Issue - State: closed - Opened by DanielRoeder1 about 1 year ago - 3 comments

#851 - Support merge lora module for 4bit and 8bit linear

Pull Request - State: closed - Opened by jiqing-feng about 1 year ago - 8 comments

#851 - Support merge lora module for 4bit and 8bit linear

Pull Request - State: closed - Opened by jiqing-feng about 1 year ago - 8 comments

#850 - Bug when loading modules_to_save and disabling adapters

Issue - State: closed - Opened by BenjaminBossan about 1 year ago - 3 comments
Labels: bug

#850 - Bug when loading modules_to_save and disabling adapters

Issue - State: open - Opened by BenjaminBossan about 1 year ago
Labels: bug

#849 - Release version 0.6.0.dev0

Pull Request - State: closed - Opened by pacman100 about 1 year ago - 1 comment

#849 - Release version 0.6.0.dev0

Pull Request - State: closed - Opened by pacman100 about 1 year ago - 1 comment

#848 - DOC: Add a contribution guide

Pull Request - State: closed - Opened by BenjaminBossan about 1 year ago - 2 comments

#847 - Fixed typos in custom_models.mdx

Pull Request - State: closed - Opened by Psancs05 about 1 year ago - 1 comment

#846 - Error in Prefix-tuning

Issue - State: closed - Opened by suntea233 about 1 year ago - 5 comments

#845 - Are the layers specified in "modules_to_save" not saved in the checkpoints during the training?

Issue - State: closed - Opened by yutanozaki1 about 1 year ago - 5 comments

#844 - Fix typos in ia3.py

Pull Request - State: closed - Opened by metaprotium about 1 year ago - 5 comments

#843 - Feature Request: 4-bit IA3 Support (comparison benchmarks included)

Issue - State: closed - Opened by His-Wardship about 1 year ago - 2 comments

#842 - Support 4-bit bitsandbytes models

Issue - State: closed - Opened by johnowhitaker about 1 year ago - 2 comments

#841 - relora

Issue - State: closed - Opened by thistleknot about 1 year ago - 4 comments

#841 - relora

Issue - State: open - Opened by thistleknot about 1 year ago

#840 - Type annotation fix

Pull Request - State: closed - Opened by vwxyzjn about 1 year ago - 1 comment

#840 - Type annotation fix

Pull Request - State: closed - Opened by vwxyzjn about 1 year ago - 1 comment

#839 - clarify the new model size

Pull Request - State: closed - Opened by stas00 about 1 year ago - 1 comment

#838 - Loading peft model with special tokens

Issue - State: open - Opened by Lathashree01 about 1 year ago - 1 comment

#838 - Loading peft model with special tokens

Issue - State: closed - Opened by Lathashree01 about 1 year ago - 2 comments

#837 - Not Able to Create and Train Multiple LoRA Adapters

Issue - State: open - Opened by DefinitlyEvil about 1 year ago - 8 comments

#837 - Not Able to Create and Train Multiple LoRA Adapters

Issue - State: closed - Opened by DefinitlyEvil about 1 year ago - 12 comments