Resources

大佬有没有兴趣再搞一搞llama3-8b

#24 opened 23 days ago by

AlexLee01

Hello

#23 opened 2 months ago by

huangfeilong

model = model.half().quantize(4).cuda() 运行显示错误"AttributeError: 'Linear' object has no attribute 'bias'"

#22 opened 2 months ago by

Frank1983823

请问训练所用的数据集能否公开？

#21 opened 2 months ago by

sssssimeng

请问作者，rlhf的actor loss是否下降和正常收敛呢？能不能给一些经验的超参数设置？请教了

#20 opened 6 months ago by

hepansls

rlhf

#19 opened 7 months ago by

nlpx

Cannot copy out of meta tensor; no data!，报错代码地方为model = model.half().quantize(4).cuda() ，猜测是量化相关问题或者作者的模型上传的时候有遗漏的文件

#18 opened 7 months ago by

yoma0101

关于两种加载模型文件方式的区别

#17 opened 8 months ago by

rk686

如何在多gpu上加载

#16 opened 8 months ago by

jersonal

关于模型权重

#15 opened 9 months ago by

mzhe

Lora和RLHF训练的代码开源了吗

#14 opened 9 months ago by

tenghg

最大token

#12 opened 9 months ago by

goodboys

还可以再次进行自我认知的lora的训练吗

#11 opened 9 months ago by

goodboys

api 调用您的模型出现错误

#10 opened 9 months ago by

neteasy

牛的无限长的原理和chatglm2是一个道理吗？

#9 opened 9 months ago by

szu2018chenli

测试了一下，很好用，比chatglm2还好用。期待更多作品。

#8 opened 9 months ago by

alfgo

训练代码

#6 opened 10 months ago by

xiazhentao

启动加载很慢，需要130秒

#5 opened 10 months ago by

devillaws

的确很好用！！

#4 opened 10 months ago by

lii1314520

会不会基于ChatGLM2-6B进行迭代？

#3 opened 10 months ago by