Issues
- 2
请问如何从头开始训练模型
#13 opened by 17702513221 - 3
- 4
- 7
论文问题: The GPT2-chitchat reaches the highest distinct scores but poor generation quality where we attribute it to the small scale of the model.
#53 opened by Ultraman-Orb - 2
- 14
在fine-tune开始时,程序提示没有tokenizer.json对训练结果有影响吗?
#48 opened by Ultraman-Orb - 1
- 1
- 9
请问有数据清洗部分的代码吗?
#29 opened by lightcome - 4
关于计算评价指标
#56 opened by huanghonggit - 6
- 7
[speaker1],[speaker2] BertTokenizer 问题
#57 opened by wsp317 - 1
Bleu-2问题
#61 opened by Ultraman-Orb - 2
加入电商对话数据集E-commerical Conversation Corpus的考量
#60 opened by BeyondSelves - 1
对比试验
#58 opened by Ultraman-Orb - 4
json配置文件问题
#59 opened by Ultraman-Orb - 7
为什么模型输入不用包括attention_mask呢?
#43 opened by li3cmz - 2
Noam scheduler 的 lr
#52 opened by Chiyu-Song - 6
验证集的ppl代码
#54 opened by Ultraman-Orb - 3
loss = lm_loss / int(args.gradient_accumulation_steps) TypeError: unsupported operand type(s) for /: 'str' and 'int'
#47 opened by Ultraman-Orb - 1
评价指标问题:请问在微调中的'average_ppl'是readme里的ppl吗?
#51 opened by Ultraman-Orb - 2
- 2
[speaker1]/[speaker2]并没有被正确编码?
#44 opened by axzz - 1
- 5
ppl 困惑度越来越大问题
#49 opened by Ultraman-Orb - 28
关于如何调用微调后的模型有些疑问
#34 opened by 27182812 - 1
用自己的语料训练时出现index out of range
#32 opened by BFJL - 1
- 8
请问你们在LCCC pretrained后,在LCCC数据集上的指标结果是怎样的呢?
#41 opened by li3cmz - 2
您好,请问LCCC-base的train,valid, test是怎么划分的
#46 opened by BeyondSelves - 2
请问有没有LCCC数据集上的PPL指标
#39 opened by iseesaw - 0
求助与文件结构
#37 opened by tongchangD - 2
关于WBDataset中标签生成的问题
#35 opened by Zessay - 6
关于 STC 微调复现效果的疑问
#33 opened by czwlines - 4
Question on the comparison between GPT and GPT2
#28 opened by mrzjy - 6
- 4
预测lccc数据集差异较大问题
#30 opened by wuzhiye7 - 1
- 1
请问,fine tuning 的时候,学习率调整为多少比较合适?谢谢
#25 opened by wulaoshi - 6
Can't download GPT_LCCC-large.zip
#24 opened by williamchai - 3
在不同机器训练和交互的问题
#23 opened by coranholmes - 5
How to select the most appropriate response with some ranking strategies ?
#18 opened by huangdacheng - 3
- 4
关于自定义数据微调实验的疑问
#22 opened by coranholmes - 6
STC的微调实验
#15 opened by JansonKong - 1
Discussion about method of self-attention.
#17 opened by xiejiachen - 1
The main difference with chitchat
#16 opened by xiejiachen - 3
超过512字符未自动进行截断
#11 opened by natureLanguageQing - 1
请问识别速度如何,可以用来作为语音识别的语言模型吗
#9 opened by 17702513221 - 2
STC微调事项
#10 opened by JansonKong