yongzhuo/ChatGLM2-SFT

Issues

• 保存模型save checkpoint怎么只保存adapter model

  #5 opened a year ago by TingJShen
  1

• 请问指定了epoch的数量，但是训练过程中打印的loss里显示epoch没有全部训练完，进度条却走完了，请问这是什么情况呀

  #12 opened a year ago by lyt719
  1

• 有试过使用 chatglm2 实现 RLHF 吗

  #11 opened a year ago by diaojunxian
  1

• 有多卡版本么

  #9 opened a year ago by jby20180901
  1

• OutOfMemoryError: CUDA out of memory.

  #10 opened a year ago by logcy
  1

• 预测程序运行，预测的都是错的

  #8 opened a year ago by xbsdsongnan
  9

• 接口没反应

  #7 opened a year ago by xbsdsongnan
  2

• RuntimeError: torch.cat(): expected a non-empty list of Tensors

  #2 opened a year ago by mathCrazyy
  1

• IndexError: piece id is out of range.

  #3 opened a year ago by ccdf1137
  1

• 你好，请教一下[GMASK]在其中的作用是什么

  #6 opened a year ago by QJShan
  1

• 如何多卡训练

  #4 opened a year ago by Sdamuu
  1

• 推理中报错：ValueError

  #1 opened a year ago by majortsui
  7