hiyouga/FastEdit

Issues

• 更大模型进行编辑会报OOM问题

  #30 opened 3 months ago by echo-valor
  0

• 尝试了下baichuan13b，似乎效果不是很好

  #29 opened 4 months ago by tianmala
  0

• It seems like there is a ignored value in delta calculation?

  #28 opened 6 months ago by rikoras
  3

• Why modifying down_proj in llama?

  #5 opened a year ago by Wenyueh
  1

• [Llama-2-7b-chat] RuntimeError: expected scalar type Float but found Half

  #27 opened a year ago by tczbzb
  7

• 使用在线量化的baichuan 13b chat 报错 LookupError: model.layers.5.mlp.down_proj.weight

  #14 opened a year ago by allendred
  3

• Llama-2-7b-chat - RuntimeError: Inference tensors cannot be saved for backward

  #26 opened a year ago by tczbzb
  2

• 训练方式和LLaMA-Efficient-Tuning-main区别

  #25 opened a year ago by yiGKTOL
  0

• 请问编辑后的模型储存在哪里了

  #24 opened a year ago by mumuyeye
  3

• 运行时报错

  #23 opened a year ago by mumuyeye
  2

• RuntimeError: computing v Vector

  #22 opened a year ago by MaximIfergan
  0

• qwen support

  #21 opened a year ago by zlh1992
  1

• Would you consider supporting ChatGLM2-6B?

  #10 opened a year ago by zhuam
  5

• 显存占用

  #20 opened a year ago by TJSL0715
  1

• this is not a good idea, may lead to severe overfitting.

  #19 opened a year ago by dhrhzz
  1

• LLaMA-2-7b-chat Editing failed

  #18 opened a year ago by MIracleyin
  0

• 数据集格式

  #17 opened a year ago by TJSL0715
  0

• 编辑完的baichuan-13b该如何保存

  #11 opened a year ago by sichehu
  3

• Is there any way to apply this interesting algorithm to the chatGLM-6B or chatGLM2-6B models?

  #3 opened a year ago by rayoa
  2

• 单张80G卡编辑7B模型 报显存不足 想请教一下如何单机多卡去run

  #16 opened a year ago by zhangfan-algo
  2

• 这种编辑的方式有副作用吗？比如模型遗忘问题

  #15 opened a year ago by runvyang
  2

• A little mistake in HyperParams

  #13 opened a year ago by canglincuizhu
  1

• Error occurs when editing Baichuan-13B

  #7 opened a year ago by hiyouga
  2

• 错误 ：TypeError: can't convert cuda:0 device type tensor to numpy.

  #9 opened a year ago by wadang
  1

• 编辑baichuan13b的时候报错NotImplementedError

  #8 opened a year ago by sichehu
  0

• 请教如何配置config？

  #6 opened a year ago by songbaiTalk
  1

• 初步看起来，这个的原理是，捕捉到两个data在模型内部的参数diff？

  #4 opened a year ago by guotong1988
  0