new TRL comes with GRPO & MPO support for vision language models 💥



we also dropped an explainer on them & how to train with them
VSN3.39%
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 9
  • Repost
  • Share
Comment
0/400
LiquidationKingvip
· 5h ago
Who hasn't trained a few large models? What is there to talk about?
View OriginalReply0
TxFailedvip
· 21h ago
tbh saved a few gpus from melting this time ngl
Reply0
Blockblindvip
· 08-10 01:45
The trl trap has gotten bigger and bigger.
View OriginalReply0
OldLeekMastervip
· 08-08 22:17
Here it comes, this upgrade is quite powerful.
View OriginalReply0
fren.ethvip
· 08-07 20:57
The new feature is reliable and has no issues!
View OriginalReply0
MondayYoloFridayCryvip
· 08-07 20:50
Do we have to go through this too? I can't take it anymore.
View OriginalReply0
MeaninglessApevip
· 08-07 20:48
All day long just knowing to do these, is it interesting?
View OriginalReply0
UncleWhalevip
· 08-07 20:41
I feel like money is coming.
View OriginalReply0
DaoDevelopervip
· 08-07 20:33
time to dig into that grpo/mpo impl tbh
Reply0
View More
Trade Crypto Anywhere Anytime
qrCode
Scan to download Gate app
Community
English
  • 简体中文
  • English
  • Tiếng Việt
  • 繁體中文
  • Español
  • Русский
  • Français (Afrique)
  • Português (Portugal)
  • Bahasa Indonesia
  • 日本語
  • بالعربية
  • Українська
  • Português (Brasil)