yule-BUAA/MergeLM

Issues

• 7B 模型合并问题

  #41 opened 2 months ago by Garmin-Qian
  4

• Question about experimental settings

  #39 opened 2 months ago by songjhPKU
  1

• Environment issues

  #40 opened 2 months ago by harmlessSR
  4

• WizardMath-7b和WizardLM-7b模型合并问题

  #18 opened 10 months ago by sasgkhgw
  3

• python enviornment

  #35 opened 2 months ago by 1998v7
  2

• When getting Figure3, and 4, you apply DARE on all layer seperately?

  #38 opened 2 months ago by SpeeeedLee
  3

• about Table1

  #37 opened 2 months ago by LianShuQuan
  3

• alpaca eval 评测问题

  #34 opened 3 months ago by qq31415926
  6

• Why same values of evaluation metrics when run>=1

  #36 opened 3 months ago by zwxu064
  2

• Any solution to merge 3B models?

  #32 opened 3 months ago by KADCA21
  1

• Are the classification heads merged?

  #31 opened 3 months ago by SpeeeedLee
  1

• Alpaca_eval evaluation error

  #33 opened 3 months ago by wang-kee
  2

• 如何合并多个模型？

  #29 opened 4 months ago by guanfaqian
  19

• 是否可以Drop 100%？？？

  #30 opened 4 months ago by Guozhenyuan
  6

• Script to reproduce all experiment in paper

  #28 opened 4 months ago by pkulium
  1

• Seeking mirrors of WizardLM models

  #27 opened 5 months ago by Sophia118Guo
  3

• Questions about randomly set delta parameters==zero

  #26 opened 7 months ago by Fay-why
  2

• AssertionError: cannot find file trainer_state.json!

  #25 opened 8 months ago by Synnai
  10

• Couldn't find a dataset script at /home/dell7960/PycharmProjects/DARE/MergeLM/glue/glue.py or any data file in the same directory.

  #23 opened 8 months ago by Synnai
  1

• ValueError: BuilderConfig 'rte' not found. Available: ['default']

  #24 opened 8 months ago by Synnai
  2

• 模型支持

  #22 opened 9 months ago by clclclaiggg
  2

• 如何对齐论文中 LM&Math&Code融合的指标

  #21 opened 9 months ago by RDXSun
  4

• 使用merge_llms_instruct_math_code.py在评估Math数据集的时候CUDA out of memory

  #20 opened 9 months ago by RDXSun
  2

• PEFT integration of DARE method

  #19 opened 10 months ago by pacman100
  1

• WizardCoder-Python-7B模型精度问题

  #17 opened a year ago by llwx593
  4

• 使用ties和magnitude方法遇到了一些问题

  #16 opened a year ago by HypherX
  1

• 你好，关于融合code模型的选择问题

  #15 opened a year ago by anon6662
  1

• The question about encoder-based model merge.

  #13 opened a year ago by bestfleer
  2

• 为什么融合wizard-lm和math后模型生成乱码

  #14 opened a year ago by anon6662
  4

• WizardMath model embedding层维度问题。

  #12 opened a year ago by sasgkhgw
  2

• Does this work with Encoder-Decoder models like T5?

  #11 opened a year ago by Inc0mple
  2

• 看起来参数是随机丢弃了

  #10 opened a year ago by ReactiveCJ
  2

• Is it possible to merge 2 LLAMA derive models like LLaVA and CodeLLAMA?

  #9 opened a year ago by ramkumarkoppu
  2

• What is the LICENSE type of this repo?

  #8 opened a year ago by ramkumarkoppu
  2

• Is there merged model available for download?

  #6 opened a year ago by kexul
  9

• Get NULL Output After Dropout w/wo Rescale

  #4 opened a year ago by LZY-the-boys
  4

• Llama model support

  #5 opened a year ago by paulcx
  6

• Can models of different architectures be merged?

  #7 opened a year ago by win10ogod
  2

• 关于delta权重

  #3 opened a year ago by zyh3826
  2

• Is the environment right? vllm 0.11.4

  #1 opened a year ago by tianyumyum
  2