thu-coai/CDial-GPT

Issues

• 预测问题

  #99 opened 2 years ago by terminator123
  3

• 支持外文吗

  #114 opened 9 months ago by srfuture
  0

• 这个tokenizer 分词有点慢啊

  #105 opened 2 years ago by minmummax
  1

• training cuda error

  #65 opened 4 years ago by Cherryjingyao
  3

• nll_loss_forward_reduce_cuda_kernel_2d: block: [0,0,0], thread: [18,0,0] Assertion t >= 0 && t < n_classes failed.

  #113 opened 2 years ago by erminga
  0

• lm_loss类型是str，无法计算loss

  #110 opened 2 years ago by ziqing0701
  2

• 模型维度问题

  #100 opened 2 years ago by terminator123
  5

• 预训练模型问题

  #111 opened 2 years ago by Junlong-Wang
  0

• finetune时的log信息在find CDial-GPT_LCCC-large里面的几个json文件，但download下来的CDial-GPT_LCCC-large里面并没有这几个文件，最后报错，报错信息如下，请问是哪里出了问题

  #108 opened 2 years ago by haiqizhang
  1

• The size of tensor a (571) must match the size of tensor b (512) at non-singleton dimension 3.

  #88 opened 3 years ago by jiangliqin
  2

• 请问在自己的数据上所有的词都用‘ ’（空格）分开吗？

  #109 opened 2 years ago by Tylerjoe
  0

• 有关预训练模型的回复结果的问题

  #107 opened 2 years ago by svjack
  1

• RuntimeError: CUDA out of memory.

  #101 opened 2 years ago by Deerzh
  12

• 请问这个微博原始数据Weibo Corpus有在哪里提供吗？可以分享吗？

  #106 opened 2 years ago by RXJ588
  0

• nll_loss_forward_reduce_cuda_kernel_2d: block: [0,0,0], thread: [18,0,0] Assertion t >= 0 && t < n_classes failed.

  #81 opened 3 years ago by Alexia1994
  4

• IndexError: Target -1 is out of bounds.

  #104 opened 2 years ago by guantao18
  0

• seeking STC dataset

  #98 opened 2 years ago by Zhou-Zoey
  3

• RuntimeError: Address already in use

  #103 opened 2 years ago by AlexKai1
  2

• 支持incremental decoding吗？

  #93 opened 2 years ago by songmzhang
  3

• _pickle.UnpicklingError: invalid load key, 'v'.

  #94 opened 3 years ago by cuiding
  3

• 请问下GPT_{Novel}模型在哪里可以找到呢？

  #96 opened 2 years ago by mianzhang
  1

• 启用半精度后，训练loss = nan

  #102 opened 2 years ago by WuDiDaBinGe
  0

• RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`

  #70 opened 3 years ago by RyanYip-Kat
  2

• 训练资源和时间没那么充裕，可以提供下跑完的模型结果吗

  #69 opened 3 years ago by jiangliqin
  4

• 学习率的问题?学习率最大6.25e-5

  #97 opened 2 years ago by WuDiDaBinGe
  2

• 评估指标ppl

  #95 opened 3 years ago by xiao-ming-code
  0

• 使用新版本 transformers[4.8.2] load CDial-GPT2_LCCC-base 模型报错

  #68 opened 3 years ago by wsp317
  3

• train

  #73 opened 3 years ago by xiao-ming-code
  1

• 用TFOpenAIGPTLMHeadModel.from_pretrained(……,from_pt=True)加载pytorch模型效果不佳

  #75 opened 3 years ago by qiuxia-alone
  1

• from pretrian error

  #64 opened 3 years ago by iris-qq
  1

• TypeError: string indices must be integers

  #72 opened 3 years ago by Jyukai02
  2

• 您好，是否方便提供数据清洗中标注数据？

  #71 opened 3 years ago by wakafengfan
  1

• interact为什么会出现后面无限循环的情况？如何解决呢？

  #90 opened 3 years ago by jiangjyjy
  1

• LCCC new link

  #92 opened 3 years ago by stephenroller
  3

• infer

  #76 opened 3 years ago by xiao-ming-code
  1

• 训练的时候数据

  #77 opened 3 years ago by xiao-ming-code
  1

• interact.py 错误

  #84 opened 3 years ago by Amazing-J
  1

• 关于预训练时小说数据的格式问题

  #80 opened 3 years ago by realLINorth
  5

• 用STC数据进行fine-tune是，用哪个指标保存最好的模型

  #83 opened 3 years ago by GongMingGithub
  2

• 云盘无法下载

  #91 opened 3 years ago by qiangqiang-he
  1

• embedding average 计算中中文分词是如何处理的。

  #87 opened 3 years ago by allyouneeds
  1

• 直接用中文文本训练

  #89 opened 3 years ago by nuochenpku
  3

• evaluation

  #85 opened 3 years ago by zhao1402072392
  0

• embedding average 计算中ground truth中存在英文，英文分词是如何处理的。

  #86 opened 3 years ago by allyouneeds
  2

• LCCC-base-split不能解压

  #82 opened 3 years ago by bingo789
  1

• 能不能公开一下LCCC-base未分词的数据集

  #79 opened 3 years ago by leelinglin
  1

• 你们预训练使用的中文小说数据在哪能下载？

  #78 opened 3 years ago by demeiyan
  1

• 结果并没有中介绍的好

  #74 opened 3 years ago by changleilei
  2

• 字嵌入向量问题

  #67 opened 3 years ago by Ultraman-Orb
  1

• 关于评价指标的问题

  #63 opened 4 years ago by AKkellan
  0