Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / stanfordnlp/pyreft issues and pull requests

#160 - [P1] customize a reward model

Issue - State: open - Opened by XiaoMaGe-hero 6 days ago - 1 comment

#159 - [P1] Add resume checkpointing support (#151)

Pull Request - State: closed - Opened by frankaging 11 days ago

#158 - [P1] Code refactory

Issue - State: open - Opened by frankaging 12 days ago
Labels: help wanted

#157 - [P1] How to load a model from ReftTrainerForCausalLM checkpoint

Issue - State: closed - Opened by weissercn 12 days ago - 2 comments
Labels: enhancement

#155 - [P0] Fixing trainer saving due to FSDP integration (#154)

Pull Request - State: closed - Opened by frankaging 13 days ago

#153 - [Minor] Update setup.py

Pull Request - State: closed - Opened by frankaging 14 days ago

#152 - [P0] Enabling FSDP with pyvene backbone change

Pull Request - State: closed - Opened by frankaging 15 days ago

#151 - [P1] How to resume checkpoint?

Issue - State: closed - Opened by mrsempress 27 days ago - 1 comment
Labels: enhancement

#150 - [P1] Weird Memory issues during finetuning

Issue - State: open - Opened by LuisMalhadas about 1 month ago
Labels: question

#148 - [P0] Fix ReftSupervisedDataset for composition

Pull Request - State: closed - Opened by PinetreePantry about 2 months ago

#147 - [P1] AttributeError: 'CausalLMOutputWithPast' object has no attribute 'mean'

Issue - State: open - Opened by krishnardt about 2 months ago - 1 comment
Labels: enhancement

#146 - Fix model load for unsupported models by pyvene

Pull Request - State: open - Opened by Masao-Taketani 2 months ago

#145 - [P2] TypeError: Object of type type is not JSON serializable

Issue - State: closed - Opened by krishnardt 2 months ago - 3 comments
Labels: question

#144 - [P0] Integration with FSDP

Issue - State: closed - Opened by xlxcomputer 2 months ago - 2 comments
Labels: help wanted, engineering

#143 - Add example Multi-GPU training script using pyreft

Pull Request - State: closed - Opened by ramvenkat98 3 months ago

#142 - [P2] Model responses repeat a lot

Issue - State: open - Opened by mfirth-truffle 3 months ago - 1 comment
Labels: question

#141 - [P0] Can't reproduce commonsense reasoning numbers

Issue - State: closed - Opened by dyahadila 4 months ago - 6 comments

#140 - RuntimeError: Numpy is not available

Issue - State: closed - Opened by gdevakumar 4 months ago

#139 - [P1] Question regarding training flag.

Issue - State: open - Opened by m-dev12 4 months ago - 4 comments
Labels: question

#138 - My mistake. Solved.

Issue - State: closed - Opened by heroz-mtaketani 4 months ago

#137 - Support Llama 3.1

Issue - State: closed - Opened by lyy1994 5 months ago - 1 comment

#136 - [P1] GPU Memory usage issue

Issue - State: open - Opened by TranscenderNing 5 months ago - 2 comments
Labels: question

#135 - [P1] Unable to access gradients of intervention blocks

Issue - State: closed - Opened by m-dev12 5 months ago - 1 comment
Labels: question

#134 - [Minor] fix undefined var

Pull Request - State: closed - Opened by frankaging 6 months ago

#133 - [Minor] Update notebook with newer names (#132)

Pull Request - State: closed - Opened by frankaging 6 months ago

#132 - Name 'LearnedSourceLowRankRotatedSpaceIntervention' is not defined in tutorial notebook

Issue - State: closed - Opened by Z1zs 6 months ago - 1 comment
Labels: bug, question

#131 - Check in negated reft notebook

Pull Request - State: closed - Opened by PinetreePantry 6 months ago

#130 - Debug subspace composition notebook impl

Pull Request - State: closed - Opened by PinetreePantry 6 months ago

#129 - [P1] Trouble Getting a ReFT Model with Pythia

Issue - State: closed - Opened by ShayanShamsi 6 months ago - 1 comment
Labels: question

#128 - [P1] Addressing issues in DPO Training (#127)

Pull Request - State: closed - Opened by AmirZur 6 months ago - 1 comment

#127 - [P1] About DPO training

Issue - State: closed - Opened by HongzhengYang 6 months ago - 3 comments

#126 - [P1] Saving/loading issues

Issue - State: open - Opened by aryamanarora 7 months ago

#125 - [P1] Cannot load trained model anymore - "type must be tuple of ints,but got NoneType"

Issue - State: open - Opened by chris-aeviator 7 months ago - 9 comments
Labels: question

#124 - [P1] Reproducing commonsense results

Issue - State: closed - Opened by jacknewsom 7 months ago - 2 comments
Labels: question

#123 - [P0] Fixing LoReFT rotation layer hot loading problem (#114)

Pull Request - State: closed - Opened by frankaging 7 months ago - 2 comments

#122 - [P1] Intervention Locations more than Prefix and Suffix

Issue - State: open - Opened by comeandcode 7 months ago - 5 comments
Labels: question

#121 - [P1] Support for non-HuggingFace models?

Issue - State: closed - Opened by 5hadytru 7 months ago - 1 comment
Labels: question

#119 - Support VLBart with ReFT

Pull Request - State: open - Opened by PinetreePantry 7 months ago

#118 - Merge "main" into peterwz-dora

Pull Request - State: closed - Opened by PinetreePantry 7 months ago

#117 - Different Intervention Locations on Different Model Components

Issue - State: closed - Opened by comeandcode 7 months ago - 2 comments

#116 - [P1] Contact information

Issue - State: closed - Opened by automateyournetwork 8 months ago - 1 comment
Labels: question

#115 - [P1] Error(s) in loading state_dict for Linear

Issue - State: open - Opened by Hamana0509 8 months ago - 2 comments
Labels: question

#113 - [P1] Getting key error in parameter while training REFT using LLAMA3

Issue - State: open - Opened by AkashGhosh 8 months ago - 9 comments
Labels: question

#112 - [P1] Loading ReFT for Llama3 model after fine-tuned with ReFT and LoRA

Issue - State: closed - Opened by Hamana0509 8 months ago - 4 comments
Labels: question

#110 - [P1] For left_padding in compute_metrics.py

Issue - State: open - Opened by mrsempress 8 months ago - 2 comments
Labels: question

#109 - [P1] Refactor ReftTrainer to save artifacts with the config

Issue - State: open - Opened by BryanWBear 8 months ago - 1 comment
Labels: enhancement, engineering

#107 - [P1] Experimental setup for instruction following experiments in the ReFT paper

Issue - State: closed - Opened by savadikarc 8 months ago - 3 comments
Labels: question

#106 - [P1] TypeError: Object of type type is not JSON serializable

Issue - State: closed - Opened by ajayspatil7 8 months ago - 4 comments
Labels: question

#105 - [P1] Possible to do batch inference?

Issue - State: open - Opened by thistleknot 8 months ago - 3 comments
Labels: question

#104 - [Major][pyreft-core] ReFT next release items

Issue - State: open - Opened by frankaging 9 months ago - 1 comment
Labels: enhancement, engineering

#103 - [P0] Revert back to ortho init as unstable training

Pull Request - State: closed - Opened by frankaging 9 months ago

#102 - [P1] [Error] can not use bfloat16 and TypeError: Object of type type is not JSON serializable

Issue - State: closed - Opened by mrsempress 9 months ago - 22 comments
Labels: question

#100 - [Minor] Basic support of quantization

Pull Request - State: closed - Opened by frankaging 9 months ago

#99 - [P1] Is it possible to merge the base model + REFT model into only model?

Issue - State: closed - Opened by celsowm 9 months ago - 1 comment
Labels: question

#98 - [P1] Loss decrease slow in readme demo when use NousResearch/Llama-2-7b-chat-hf

Issue - State: closed - Opened by svjack 9 months ago - 2 comments
Labels: question

#97 - [P0] Does this project support turning in 4bit or 8bit Quantify？

Issue - State: closed - Opened by svjack 9 months ago - 5 comments
Labels: enhancement, engineering

#96 - [P1] Multiple Positions Intervention

Issue - State: closed - Opened by comeandcode 9 months ago - 1 comment
Labels: question

#95 - [P1] Questions on differences between paper and code

Issue - State: closed - Opened by calpt 9 months ago - 2 comments
Labels: question

#94 - [Minor] Enable lora with loreft training

Pull Request - State: closed - Opened by frankaging 9 months ago

#93 - [P1] support ReFT+PEFT by using ReftModel to wrap PeftModel (#46)

Pull Request - State: closed - Opened by frankaging 9 months ago - 1 comment

#92 - [P1] Transitioning from peft to pyreft for Classification Approach

Issue - State: open - Opened by SaBay89 9 months ago - 2 comments
Labels: question

#91 - [P1] Model Compatibility

Issue - State: closed - Opened by SaBay89 9 months ago - 2 comments
Labels: question

#90 - forward() got an unexpected keyword argument 'unit_locations'

Issue - State: closed - Opened by xerkey 9 months ago - 2 comments

#87 - [P1] TypeError: train() takes 1 positional argument but 2 were given

Issue - State: closed - Opened by alpozdarendeli 9 months ago - 1 comment
Labels: question

#86 - [P1] Loading REFT fro RoBERTa Models

Issue - State: open - Opened by hSterz 9 months ago - 4 comments
Labels: question

#85 - [P0] Make `make_last_position_supervised_data_module` parallelizable to speed up processing!

Issue - State: open - Opened by truskovskiyk 9 months ago - 2 comments
Labels: enhancement

#84 - [P1] Convert reft model to hf model

Issue - State: closed - Opened by thu-yn 9 months ago - 1 comment
Labels: question

#83 - [P1] Getting error as IntervenableModel.train() takes 1 positional argument but 2 were given

Issue - State: closed - Opened by atharvapatiil 9 months ago - 4 comments
Labels: question

#81 - [P1] How did you create the validation set for Commonsense reasoning hyperparameter tuning?

Issue - State: closed - Opened by Edenzzzz 9 months ago - 5 comments
Labels: question

#80 - Getting issue while loading Phi3 in reft_model

Issue - State: closed - Opened by atharvapatiil 10 months ago - 9 comments

#79 - [P1] RuntimeError: cutlassF: no kernel found to launch!

Issue - State: closed - Opened by ds-praveenkumar 10 months ago - 4 comments
Labels: question

#78 - [P1] catastrophic forgetting

Issue - State: closed - Opened by jiacheo 10 months ago - 1 comment
Labels: question

#77 - [P1] Intuitive-wise, should we keep the projection orthogonal during training?

Issue - State: closed - Opened by Edenzzzz 10 months ago - 2 comments
Labels: question

#76 - ReFT + DPO Tutorial

Pull Request - State: closed - Opened by AmirZur 10 months ago - 1 comment

#75 - [Minor] fix subspace (#72)

Pull Request - State: closed - Opened by frankaging 10 months ago - 1 comment

#74 - [Minor] More refactory to support Llama3 experiments

Pull Request - State: closed - Opened by frankaging 10 months ago

#73 - [P1] Confirmation of alpaca_eval version

Issue - State: closed - Opened by BaohaoLiao 10 months ago - 4 comments
Labels: question

#72 - [P0] compreft.ipynb error = KeyError: 'subspaces'

Issue - State: closed - Opened by RonanKMcGovern 10 months ago - 5 comments
Labels: bug

#71 - [P1] Location of code for "LM training and serving with ReFT"

Issue - State: open - Opened by RonanKMcGovern 10 months ago - 2 comments
Labels: enhancement

#70 - [P2] Pyreft tensorboard integration

Issue - State: open - Opened by PinetreePantry 10 months ago
Labels: bug

#69 - [P1] TypeError: Object of type type is not JSON serializable

Issue - State: closed - Opened by srn-source 10 months ago - 7 comments
Labels: question

#67 - [P0] Why is the number of trainable parameters for prefix-tuning is 0.11%

Issue - State: closed - Opened by BaohaoLiao 10 months ago - 7 comments
Labels: documentation, question

#66 - [P0] Adding DPO Support

Issue - State: closed - Opened by jinzhuoran 10 months ago - 9 comments
Labels: enhancement, help wanted

#65 - [P1] I am bit confused how to reproduce Table 2 (all baselines + main method)

Issue - State: closed - Opened by sanyalsunny111 10 months ago - 3 comments
Labels: question

#64 - [Major] Support Llama3 models

Pull Request - State: closed - Opened by frankaging 10 months ago

#63 - [P1] TGI and vLLM support

Issue - State: open - Opened by RonanKMcGovern 10 months ago - 7 comments
Labels: question

#62 - [P1] MNLI has two validation set, how do you report the score

Issue - State: closed - Opened by BaohaoLiao 10 months ago - 3 comments
Labels: question

#61 - [P1] OpenAI CLIP model

Issue - State: open - Opened by sailfish009 10 months ago
Labels: enhancement

#60 - [P1] How to attend to memorized intervention?

Issue - State: open - Opened by chris-aeviator 10 months ago - 2 comments
Labels: question

#59 - [P0] How do I train more than 1 layer at a time?

Issue - State: closed - Opened by thistleknot 10 months ago - 4 comments
Labels: enhancement