yangjianxin1/GPT2-chitchat

Issues

• 训练模型问题

  #133 opened 6 months ago by AlisonDexter
  3

• dataset

  #129 opened 9 months ago by 1dengguosong
  1

• 训练结果可视化

  #131 opened 7 months ago by Jasonup6
  1

• AttributeError: 'tokenizers.Encoding' object has no attribute 'n_sequences'

  #132 opened 7 months ago by AlisonDexter
  1

• 新手求帮助啊，友友们

  #130 opened 8 months ago by dianfengxiaobo
  2

• 输入ctrl+z程序不结束，继续进行对话

  #80 opened 3 years ago by Chunfeng1994
  2

• ZeroDivisionError: division by zero

  #126 opened a year ago by Delimeng
  6

• 大佬好，我用两张3090去使用你所提供的那个100w的数据集训练，可是很快loss稳定在3.0就不降低了

  #101 opened 2 years ago by iniroc
  8

• 有没有可能通过对通用对话的模型进行fine-tune给机器人赋予一个固定的人设和状态？并且对客观条件做出正确的反应？

  #96 opened 2 years ago by lewiswu1209
  3

• 有关对预训练模型微调的问题

  #128 opened a year ago by Edisonhimself
  0

• padding mask

  #127 opened a year ago by yttas
  0

• 请问dataset里面没有传入labels,后面dataloader是从哪里得到的labels的呢

  #95 opened 3 years ago by pureblacker
  4

• train.py报错ZeroDivisionError: division by zero

  #107 opened 2 years ago by wangfan002
  6

• 是本身就存在的train.py 里的bug吗？ 写入正确的数据path，成功导入数据后仍然会报的valueerror问题 到底是哪一步错了？

  #79 opened 3 years ago by Saraooe
  12

• 想使用GPT2的微调来实现负样本的生成

  #124 opened a year ago by jazzlee008
  0

• 网盘链接失效

  #123 opened a year ago by qhykwsw
  0

• 如何將模型轉換為繁體中文版

  #103 opened a year ago by yan-930521
  1

• 下载问题

  #121 opened a year ago by Thexiaoyuqaq
  0

• 训练代码逻辑问题

  #119 opened a year ago by wujohns
  2

• 请问有办法集成到web应用里吗？

  #100 opened 2 years ago by drizzt00s
  5

• 训练时的一些错误

  #115 opened 2 years ago by gittyspirit
  1

• README中的小错误，但是可能造成不必要的麻烦

  #112 opened 2 years ago by SchweitzerGAO
  3

• 更换更大的gpt2模型进行训练 如：gpt2_large

  #94 opened 3 years ago by htthYjh
  1

• fine-tuning

  #111 opened 2 years ago by feyxong
  1

• 模型加载报错

  #93 opened 3 years ago by UnstoppableCurry
  6

• 模型层数问题

  #117 opened 2 years ago by Junlong-Wang
  0

• ignore_index 的设置

  #116 opened 2 years ago by xuyouqian
  3

• 有什么方法让已经训练好的模型去训练用户和bot的对话呢？

  #114 opened 2 years ago by smallcat9669
  0

• 请问断点续训是否是不支持的？

  #113 opened 2 years ago by mike3090
  1

• 为什么要删除mmi的方法啊？效果不好吗？

  #110 opened 2 years ago by Jehuty-ML
  0

• Uuuu

  #108 opened 2 years ago by wjbf2045
  0

• 技术交流群

  #106 opened 2 years ago by bouxin
  0

• 这是什么问题？小白学习中

  #105 opened 2 years ago by fuzhengwei
  0

• 自定义语料 报错ValueError: num_samples should be a positive integer value, but got num_samples=0

  #89 opened 3 years ago by Dynamicboboo
  2

• 为什么训练时候没有设定attention_mask这个参数

  #87 opened 2 years ago by Choitsugun
  1

• pytorch_model.bin如何部署到android手机上？

  #102 opened 2 years ago by cxhermagic
  0

• 请问预训练参数去哪里下载啊？

  #97 opened 2 years ago by 13227721183
  5

• 为什么我这训练过程中学习率是逐步上升的啊？

  #99 opened 2 years ago by Chenny0808
  0

• gpt模型的注意力头数n_head会影响模型在哪些方面的性能？怎样有效提升模型处理多轮对话的准确率？

  #98 opened 2 years ago by oerifjmerefver
  0

• error---->train.py

  #85 opened 3 years ago by KangChou
  5

• vocab

  #90 opened 3 years ago by pureblacker
  2

• 模型训练不动了

  #91 opened 3 years ago by G-Jarvey
  5

• 预测

  #84 opened 3 years ago by zzcceeyy
  0

• 在tensorflow中使用50w chichat预模型时生成结果不佳

  #83 opened 3 years ago by qiuxia-alone
  0

• 用训练好的模型生成数据，偶尔会报下边的错，位置不一样，跟您请教一下，怎么解决呢？网上查的方法都不行

  #82 opened 3 years ago by kavinwow100
  0

• 可以加个联系方式吗大佬？

  #81 opened 3 years ago by dahefanzhou
  0

• 问一下有建议引用方式吗？

  #78 opened 4 years ago by Pwang001
  0

• 想问一下初始化的预训练模型的词表是2w+，但是现在的词表vocab.txt 只有1.3w,是根据自己的预料处理过吗？

  #77 opened 4 years ago by Cherryjingyao
  3

• 还得再请教下您temperature的用意，有点没太看懂

  #76 opened 4 years ago by kavinwow100
  1

• 拆分训练集和验证集没有再用 sklearn的train_test_split是不是因为20%的验证集数据量太大了，所以用了个常量8000

  #75 opened 4 years ago by kavinwow100
  3