Issues
- 7
bug on google colabs code
#3805 opened by View-my-Git-Lab-krafi - 4
KTO training with datasets in alpaca format
#3803 opened by Cheungki - 1
对codeqwen-base模型进行pretrain的一些问题
#3806 opened by YuMingtao0503 - 2
数据集数量加载缺失,缺失之后训练的模型输出全部为空
#3800 opened by jayofhust - 0
希望支持InternVL视觉模型的微调
#3802 opened by zhangzhongpeng02 - 1
- 0
errors while in finetune intermlm2-chat-20b with qlora
#3798 opened by a1exyu - 1
i dont want to use huggingface, i can i mention my local data path from my computer. 我不想使用Huggingface,我可以提到我电脑上的本地数据路径。
#3797 opened by View-my-Git-Lab-krafi - 2
Question: how does template work with dataset in examples: llama3_lora_sft.yaml
#3771 opened by chuangzhidan - 2
昇腾NPU使用API推理报错
#3796 opened by msqp - 1
- 1
偏好训练,如何使用ShareGPT格式数据集
#3774 opened by binganao - 0
偏好数据集 Supervised Fine-Tuning 有问题
#3793 opened by wlz987 - 8
【互助】关于Agent中react/ReAct 模板数据如何组织方法
#3750 opened by txy6666yr - 4
Sprider数据集训练,默认参数训练出的模型,需要设置提示词为Sprider中的,回答准确;自定义后的小数据集,训练过程无lass曲线,并且没有效果
#3791 opened by zhangkuo-zk - 1
bash: llamafactory-cli: command not found
#3790 opened by XuJianzhi - 3
请教全量微调时deepspeed保存事项
#3789 opened by SandroChen - 0
ascend 910b,chatglm2做全量微调报错
#3788 opened by belle9217 - 3
orpo训练较慢
#3786 opened by LiuChen19960902 - 2
SFT微调完成导出模型的API推理问题
#3787 opened by Zomun - 4
Fail to run mixture-of-depths sft official example script
#3784 opened by Zkli-hub - 1
Question about --dpo_ftx Parameter Setting
#3781 opened by zandfj - 4
在昇腾npu环境下运行报错
#3779 opened by feria-tu - 1
api部署,method not allowed问题
#3783 opened by jy-101361-1810897 - 2
请教一下如何不使用任何模版加载数据?
#3782 opened by SandroChen - 1
llama3增量预训练冻结哪些层训练哪些层效果比较好?
#3780 opened by CanvaChen - 0
adding language
#3776 opened by hglmz - 4
Rouge-score results are surprisingly low
#3764 opened by Jiminator - 1
推理阶段,预测文件中label显示不全问题
#3775 opened by jy-101361-1810897 - 5
请问量化校准数据c4_demo.json的生成有什么要求的呢?
#3773 opened by maiqingqiang - 0
How to estimate total steps and set proper
#3777 opened by zhaoxu98 - 1
[Feature Request] support for new peft model `pissa`
#3778 opened by wsp317 - 1
传lr_scheduler_kwargs参数报错error: argument --lr_scheduler_kwargs: invalid Dict value: "{'num_cycles':6}"
#3770 opened by Cucunnber - 2
训练自定义数据集出现错误
#3768 opened by zhangkuo-zk - 1
微调后的模型如何多卡推理
#3769 opened by cl12191718 - 1
在不启用流式数据读入的情况下数据是否会被shuffle
#3772 opened by Limerence0502 - 3
安装docker出现问题
#3767 opened by kynow2 - 2
lora微调后模型表现不佳 数据停止生成问题
#3766 opened by sweetning0809 - 2
postman调用api
#3759 opened by yang1111-gif - 0
自定义数据集训练的时候出现 ValueError: Expected input batch_size (103) to match target batch_size (95).
#3765 opened by keesh0410 - 2
2块A6000用lora方式训练,但是只有64G内存,在保存checkpoint时会爆内存,请问有办法减少保存checkpoint时内存使用量的设置么?
#3752 opened by Parasolation - 1
接#3740,解决不能单卡增量预训练后,多卡增量预训练出现OOM
#3757 opened by AI-X-King - 1
Orpo训练的内存占用问题
#3754 opened by MissQueen - 1
模型训练所需flops
#3763 opened by BhAem - 1
dataset_info能否支持设置固定的system值?
#3762 opened by CanvaChen - 0
如何修改预览命令,或者如何添加训练参数
#3761 opened by TC10127 - 1
关于用llama2的变种模型 做orpo微调出现tokennizer报错问题。
#3760 opened by nlpcainiao - 3
模型评价问题
#3758 opened by langlibaipiao - 4
按照README流程进行llama3 8B LORA SFT训练,得出的模型对话时不会停止,效果也一般
#3753 opened by matrixssy - 1
vllm 推理部署指令怎么调整可以在后台运行, shell关闭不会被kill
#3751 opened by lldhliu