Async RLHF Paper Checkpoints Checkpoints for "Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models" https://arxiv.org/abs/2410.18252 vwxyzjn/online_dpo_async Updated 20 days ago • 40 vwxyzjn/online_dpo_sync Updated 20 days ago • 27 vwxyzjn/ppo_async Updated 20 days ago • 23 vwxyzjn/ppo_sync Updated 20 days ago • 37
lm-human-preference-details vwxyzjn/train_policy_accelerate__sentiment_offline_5k.json__seed1__1696447674 Text Generation • Updated Oct 4, 2023 • 68 lm-human-preference-details/train_policy_accelerate__sentiment_offline_5k.json__seed1 Text Generation • Updated Oct 4, 2023 • 76
vwxyzjn/train_policy_accelerate__sentiment_offline_5k.json__seed1__1696447674 Text Generation • Updated Oct 4, 2023 • 68
lm-human-preference-details/train_policy_accelerate__sentiment_offline_5k.json__seed1 Text Generation • Updated Oct 4, 2023 • 76