Lewis Tunstall(@_lewtun):If you're using GRPO in TRL, you should really switch to the new async trainer. In our benchmarks, it's ~2-4x faster 🔥

注册并分享邀请链接，可获得视频播放与邀请奖励。

立即注册

Lewis Tunstall

@_lewtun

🤠 post-training @huggingface

加入 August 2018

549 正在关注 20.4K 粉丝

Lewis Tunstall@_lewtun

2026.08.13 13:55

If you're using GRPO in TRL, you should really switch to the new async trainer. In our benchmarks, it's ~2-4x faster 🔥

0

0

6

253

20

转发到社区

热门用户

0 粉丝

4.2M 粉丝

26.4M 粉丝

569.1K 粉丝

オリコンニュース

1.9M 粉丝

吴说区块链

@wublockchain12

181.5K 粉丝

100.9K 粉丝

ツイッター速報〜BreakingNews

256.8K 粉丝

@Kongkongda5882

34.1K 粉丝

0 粉丝

中國新聞社

547.4K 粉丝

华尔街观察 Xtrader

@cnfinancewatch

129.8K 粉丝

348.8K 粉丝

3.4M 粉丝

23.7K 粉丝