Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix PPO #2377

Merged
merged 7 commits into from
Nov 11, 2024
Merged

fix PPO #2377

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
5c90c79
fix
hjh0119 Oct 23, 2024
7968586
update
hjh0119 Nov 4, 2024
c79ece9
fix ds
hjh0119 Nov 7, 2024
7ac251e
Merge remote-tracking branch 'origin/main' into fix-1023
hjh0119 Nov 9, 2024
f5b1709
num_sample_generations
hjh0119 Nov 9, 2024
7b7c024
update
hjh0119 Nov 10, 2024
15bebfe
rm unused import
hjh0119 Nov 10, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
update
  • Loading branch information
@hjh0119
hjh0119 committed Nov 10, 2024
commit 7b7c024e62831ad66d612dbc915f1f706c0f5965
2 changes: 1 addition & 1 deletion docs/source/LLM/人类偏好对齐训练文档.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,7 +160,7 @@ RLHF中的PPO(proximal policy optimization)阶段, 涉及到四个模型
- cliprange_value: PPO价值损失函数中的clip范围, 默认为0.2
- gamma: 累计奖励的折扣因子, 默认为1.0
- lam: [GAE](https://arxiv.org/abs/1506.02438)中的lambda系数, 默认为0.95

- num_sample_generations: 训练过程中生成的调试样本数量, 默认为10

```bash
CUDA_VISIBLE_DEVICES=0 \
Expand Down
1 change: 1 addition & 0 deletions docs/source/Multi-Modal/人类偏好对齐训练文档.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
- [环境准备](#环境准备)
- [数据集](#数据集)
- [DPO](#dpo)
- [RM](#rm)
- [CPO](#cpo)
- [ORPO](#orpo)
- [SimPO](#simpo)
Expand Down
3 changes: 3 additions & 0 deletions docs/source_en/LLM/Human-Preference-Alignment-Training-Documentation.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,8 @@ This document provides training scripts for various human preference alignment a
- [Environment Setup](#environment-setup)
- [Dataset](#dataset)
- [DPO](#dpo)
- [RM](#rm)
- [PPO](#ppo)
- [KTO](#kto)
- [CPO](#cpo)
- [ORPO](#orpo)
Expand Down Expand Up @@ -159,6 +161,7 @@ Hyperparameters
- cliprange_value: Clip range in the PPO value loss function, default is 0.2
- gamma: Discount factor for cumulative rewards, default is 1.0
- lam: Lambda value for [GAE](https://arxiv.org/abs/1506.02438), default is 0.95
- num_sample_generations: Number of debugging samples generations throughout training, default is 10

```bash
CUDA_VISIBLE_DEVICES=0 \
Expand Down