microsoft/LoRA

Issues

• Understanding Figure 3

  #184 opened a month ago by Davidyao99
  0

• Pre-trained conv weight is not same as that self.conv.weight

  #108 opened a year ago by aleemsidra
  3

• LoRA adapter checkpoints not downloadable

  #141 opened a year ago by kibru9399
  9

• Reproduce Lora results is close but not accurate

  #165 opened 8 months ago by harsh306
  3

• Parameter count on GPT-2 medium

  #172 opened 7 months ago by Heimine
  1

• How are the pretrained weigts set to be fixed during training?

  #181 opened 4 months ago by HaiminZhang
  1

• NLU experiments error

  #180 opened 4 months ago by jie040109
  1

• Questions about replicating NLU experiments

  #177 opened 5 months ago by lky-violet
  0

• [Question about multi-gpu training]

  #170 opened 7 months ago by FindTheTruth
  2

• Questions about running the cola dataset script

  #176 opened 6 months ago by dengxingzhi
  1

• Not able to reproduce the scores using provided checkpoint on NLG tasks

  #138 opened a year ago by ylli0218
  5

• How to adjust LoRA into nn.ConvTranspose2d?

  #160 opened 9 months ago by vanmeruso
  2

• Where is the LoRA matrices saved?

  #173 opened 7 months ago by KelsenMa
  0

• question for scale！

  #171 opened 7 months ago by Duperr
  0

• Can't reproduce the results for GLUE and hyperparameter misalignment

  #149 opened a year ago by nbasyl
  4

• Can not reproduce the result of Roberta-Base

  #151 opened a year ago by Ther-nullptr
  2

• LORA on T5 model

  #169 opened 8 months ago by vivektreddy
  0

• Cannot use lora for a pre-trained model

  #123 opened a year ago by HelloWorldLTY
  4

• lora-dim == lora-r ?

  #168 opened 8 months ago by ardand4708
  0

• After joining Lora, the first few layers show a gradient of 0

  #167 opened 8 months ago by hluckye
  0

• Guidance Needed on Continuing Training with a New Dataset via LoRA

  #166 opened 8 months ago by TanTaiLi
  0

• Dynamic Lora Selection In Runtime❓

  #164 opened 8 months ago by teneous
  1

• _conv_forward() error

  #162 opened 8 months ago by likaiucas
  0

• Cannot implement LoRA on a custom model containing transformer encoder from pytorch

  #161 opened 9 months ago by wsuSaiman
  0

• Is the output the entire model？

  #139 opened a year ago by licy02
  3

• Layers.py not being executed

  #150 opened a year ago by Aradhye2002
  1

• Conv1d and Conv3d are not working

  #115 opened a year ago by gau-nernst
  4

• Question about the test set of the GLUE benchmark

  #145 opened a year ago by James6Chou
  1

• Is it necessary to add `model = model.merge_and_unload()` when training a new LoRA adapter?

  #156 opened 10 months ago by 4daJKong
  0

• batch=1,why adapter latency so much vs. LoRA in paper???

  #129 opened a year ago by macqueen09
  3

• how to improve the memory ability of lora fine tuning?

  #153 opened 10 months ago by K-Alex13
  1

• models are the same after loading lora parameters using peft library

  #154 opened 10 months ago by ra-MANUJ-an
  0

• How to compute that GPT-2 M (FTTop2 ) trainable parameters number is 25.19M?

  #148 opened a year ago by floatingbigcat
  0

• [Minor] Possible typos in weight initialization

  #146 opened a year ago by awgu
  0

• Question about seed numbers.

  #144 opened a year ago by SEONHOK
  0

• AttributeError: 'lora.Conv2d' object has no attribute 'weight'

  #134 opened a year ago by bjzhb666
  3

• The description and the behavior don't match

  #142 opened a year ago by yongchanghao
  2

• Question about reproducing RoBERTa base Fine-tune

  #127 opened a year ago by tlsdbfk
  1

• Support multi-lora fine tune in the same GPU

  #136 opened a year ago by merlintang
  0

• T(w) problem

  #128 opened a year ago by fclearner
  0

• Probably a bug in the lora embeding class in loralib/layers.py

  #121 opened a year ago by favorart
  2

• Some questions about LoRA for pre-trained model

  #124 opened a year ago by iopwsy
  0

• Using gradient checkpoint with LoRA

  #120 opened a year ago by dudskrk
  0

• The content on pypi does not seem to be updated

  #117 opened a year ago by MikuAndRabbit
  3

• matmul ordering in MergedLinear

  #109 opened a year ago by zhiqi-0
  1

• fine tuning RoBERTa-base with LoRA (ValueError: Classification metrics can't handle a mix of binary and multilabel-indicator targets)

  #116 opened a year ago by rozhix
  0

• How does this paper search the hyperparameters on GLUE datasets with Roberta?

  #110 opened a year ago by A11en0
  1

• Embedding reset_parameters() implement wrong

  #114 opened a year ago by chenjiasheng
  3

• PyPi version and GitHub version give different results

  #112 opened a year ago by kimihailv
  1

• RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasSgemm( handle, opa, opb, m, n, k, &alpha, a, lda, b, ldb, &beta, c, ldc)`

  #106 opened a year ago by aleemsidra
  4