OpenLMLab MOSS-RLHF issues

OpenLMLab / MOSS-RLHF

MOSS-RLHF

Apache License 2.0

1.18k stars 89 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

RM数据构造

#55 tcxia opened 1 month ago
1
Has anyone compared this training framework to TRL?

#54 StarrySeas1 opened 1 month ago
1
对第二篇论文中有些不明白的地方请教解惑

#53 Obr00007576 opened 2 months ago
0
论文中提到在PPO流程中可以固定其他模型，先训练reward model直到value loss为0，请问这边具体是怎么进行训练的呢？

#52 HCHCXY opened 2 months ago
1
Part2中meta dataset的生成

#51 yata0 opened 2 months ago
1
训练集量级咨询

#50 Macvh opened 2 months ago
1
PPOSFTDataset bug report和相关问题咨询

#49 DZ9 opened 3 months ago
1
关于rm中lm loss计算的疑问

#48 DZ9 opened 3 months ago
1
adding citation of part 2

#47 fakerbaby closed 2 months ago
0
bash train_ppo_en.sh error

#46 robotzheng closed 3 months ago
4
论文中rm对比学习训练方法疑问

#45 yhhh777 opened 3 months ago
4
Issues with using the released hh dataset.

#44 jltchiu opened 3 months ago
2
关于rm模型训练策略与损失函数

#43 tonylin52 opened 4 months ago
10
Clarification on MetaRM-optimization Implementation

#42 Benjamin-eecs opened 4 months ago
2
release the code for training the reward model

#41 refrain-wbh closed 4 months ago
0
[Question] Adaptive Margin

#40 eyuansu62 closed 4 months ago
3
请问目前支持基座模型使用Mistral-7b吗

#39 YijuGuo opened 4 months ago
1
自有的底座模型，自有的SFT权重，重新训练RM，可行么

#38 camposs1979 opened 5 months ago
1
Why are you not releasing your reward model for english?

#37 AmanSinghal927 opened 5 months ago
1
Inference with SFT and Policy EN models

#36 henrypapadatos opened 6 months ago
1
请问下代码里的kl散度问题

#35 rigorosyangffff opened 6 months ago
1
合并权重问题

#34 red-tie opened 7 months ago
6
关于reward model的权重合并问题

#33 HuipengXu opened 7 months ago
1
资源占用问题

#32 Ming-Di opened 8 months ago
3
关于reward model的部分的part 2有计划时间节点吗

#31 SpongebBob opened 8 months ago
13
Any benchmark vs SFT？

#30 guotong1988 opened 8 months ago
2
deepspeed的parameter_offload问题

#29 LiangZhuuu closed 8 months ago
1
PPO显存占用问题

#28 LiangZhuuu closed 9 months ago
0
PPO data en

#27 borisshapa opened 9 months ago
1
关于ppo阶段，reward分数计算的问题

#26 mengyanggithub opened 9 months ago
5
typo

#25 chosenone75 closed 9 months ago
1
关于中文reward-model参数合并的问题

#24 hannlp opened 9 months ago
4
关于配置环境

#23 zjutkarma closed 9 months ago
2
reward model训练的哪些方面的能力

#22 yuanhuachao opened 9 months ago
1
关于Reward model打分的一些疑惑

#21 hannlp opened 9 months ago
12
英文的PPOdata

#20 QYHcrossover opened 9 months ago
1
Training on 8 Nvidia RTX A6000

#19 Top34051 opened 10 months ago
1
value model与reward model

#18 KUANWB opened 10 months ago
2
PPO训练稳定性问题

#17 hust-kevin opened 10 months ago
5
训练reward model的脚本

#16 wangzhao88 opened 10 months ago
3
reward_model准确率

#15 mingrenbuke opened 10 months ago
1
Training script of reward model

#14 zwhe99 closed 10 months ago
2
Technical report PART 2

#13 snowkcon opened 10 months ago
3
内存占用大问题

#12 QYHcrossover closed 9 months ago
2
Reward Model

#11 Cyber-Axe opened 10 months ago
2
关于reward model

#10 skepsun closed 10 months ago
5
support lora training

#9 akk-123 closed 10 months ago
1
Can I run this pipeline on A100-40GB?

#8 zwhe99 closed 10 months ago
4
用于PPO训练的数据结构

#7 Arain-sh closed 10 months ago
2
The release of reward model training code?

#6 hejujie closed 10 months ago
3