Open-weight & owned stack · Training tuning

Hugging Face Releases TRL 1.14.0 Training Improvements

Sep 25, 2026, first seen via Training & inference tooling releases

Hugging Face released TRL 1.14.0 with streamed log-probability computation and fixes for distributed training and trainer options.

Read the original at github.comOpens the publisher's site in a new tab

More in Open-weight & owned stack

Kyutai Releases Voice Of Reason CheckpointsSep 23 Xiaomi Releases MiMo-V2.6 Open-Weight Model FamilySep 22 Intrinsic open-sources Core robotics software infrastructureSep 22 Harvey Launches Kimi-Based In-House ModelSep 21 Kimi K3 Reaches AWS And Alibaba Cloud PlatformsSep 21