beyondguo/LLM-Tuning

Issues

• chatglm2报错：ValueError: weight is on the meta device, we need a `value` to put in on 0

  #15 opened a year ago by angel1288
  3

• ImportError: dlopen: cannot load any more object with static TLS

  #59 opened 8 months ago by hjing100
  1

• Why epoch in log is different from progress

  #58 opened a year ago by jimmy-walker
  0

• 有群吗,拉个群把

  #9 opened a year ago by dragononly
  2

• ChatGLM2按照readme教程微调了，但是没有效果！！！

  #14 opened a year ago by HelixPark
  19

• 多卡训练感觉不是并发的?

  #18 opened a year ago by shenmadouyaowen
  8

• 使用ChatGLM2-6B分词报错

  #57 opened a year ago by kyle-hy
  0

• code lamma微调脚本可以使用baichuan2的吗

  #56 opened a year ago by xhaoss
  0

• 请教一个问题，chatglm2在用lora微调时，不添加attention mask也可以么？

  #21 opened a year ago by annw0922
  2

• model.hf_device_map 不存在如何解决呢

  #55 opened a year ago by LivinLuo1993
  0

• 为什么ppo model 需要接AutoModelForCausalLMWithValueHead呢？

  #51 opened a year ago by jiahuanluo
  1

• chaglm-6b lora微调执行到指定的eval_step后提示“iteration over a 0-d tensor”

  #49 opened a year ago by LivinLuo1993
  1

• 按照教程,一步一步弄的,到了训练PPO的时候, 卡到 CUDA error: device-side assert triggered

  #54 opened a year ago by karl-tao-zhang
  3

• where is "my_templates" module

  #53 opened a year ago by andyzhu
  0

• 这个微调代码不能直接用来baichuan-13B模型的微调？在13B一直报错

  #38 opened a year ago by DaiJitao
  2

• PPO training CUDA out of memory

  #52 opened a year ago by 14H034160212
  0

• chatglt-6b2 lora微调使用int4精度报错

  #50 opened a year ago by hehuomu
  0

• baichuan-13b reward model训练

  #48 opened a year ago by endlesstalking
  0

• ppo报错

  #43 opened a year ago by yyy900
  3

• baichuan_lora_tuning运行，为什么一直卡在8%呢？

  #44 opened a year ago by yanduoduan
  2

• 为什么训练阶段的显存一直在往上涨？一会就 OOM 了

  #45 opened a year ago by Amazing-J
  1

• 加载数据集报错

  #47 opened a year ago by endlesstalking
  1

• chatglm2不支持SequenceClassification请问如何解决。。

  #46 opened a year ago by jaycehw
  1

• RuntimeError: Expected is_sm80 to be true, but got false.

  #40 opened a year ago by Qiang-HU
  4

• TORCH_USE_CUDA_DSA

  #34 opened a year ago by Mrjude
  3

• 关于全量和LoRA的问题

  #28 opened a year ago by llmrainer
  2

• 如何在离线环境生成 tokenized_data ？

  #42 opened a year ago by seek4self
  1

• rulai_enhance.json 数据开源了嘛大佬

  #41 opened a year ago by zlszhonglongshen
  1

• 使用lora微调ChatGLM2-6B 报错

  #39 opened a year ago by QJShan
  1

• 如何评估finetune过程中模型的性能？

  #13 opened a year ago by dxyzx0
  2

• lora tuning 出的权重，再加一个合并的功能？

  #37 opened a year ago by litetoooooom
  1

• 微调的时候如何让模型记住一些特有的知识呢

  #29 opened a year ago by controZheng
  3

• 怎么训练多轮对话呀

  #7 opened a year ago by liuhuapiaoyuan
  5

• 6块Tesla A100 40G训练会卡住

  #35 opened a year ago by cherishtttz
  1

• datasets.builder.DatasetGenerationError: An error occurred while generating the dataset

  #36 opened a year ago by zlszhonglongshen
  1

• 请教一下显存问题

  #25 opened a year ago by Jamine-W
  3

• 是否支持多轮对话的微调？

  #33 opened a year ago by yaakua
  1

• glm2和glm分别需要多大的显存才能微调

  #32 opened a year ago by ShiXiangXiang123
  1

• cuda版本

  #30 opened a year ago by 1910183821
  1

• 请问可以有支持 Apple M 系列处理器 mps 的方式吗？

  #27 opened a year ago by minjin
  1

• 微调出错如下：

  #31 opened a year ago by ShiXiangXiang123
  1

• 是否可以增加 预测和推断的代码？

  #22 opened a year ago by RileyShe
  1

• Chinese-LLaMA-Alpaca有计划安排加入吗？

  #20 opened a year ago by SilenceWinter
  1

• 微调后输出长度不够长

  #19 opened a year ago by dragononly
  1

• weights 文件训练出来很小，正常嘛

  #16 opened a year ago by dragononly
  1

• 训练后加载跑,无效果,请看看这样对吗

  #17 opened a year ago by dragononly
  0

• chatglm2_loar_tuning run raise NotImplementedError

  #11 opened a year ago by xieyongshuai
  7

• 运行报错

  #8 opened a year ago by shenmadouyaowen
  22

• 看了#8号贴，注释了两行报119

  #12 opened a year ago by BoFan-tunning
  0

• 报错“KeyError: 'transformer.embedding'”

  #10 opened a year ago by lilulu0702
  3