Blogs

LLM Alignment using Direct Preference Optimization, an alternative to RLHF

Highlights include Llama 3, TR-DPO, Video2Game, Dynamic Typography, and new leaderboards.

Highlights of the latest release.

Highlights include Infini-attention, OSWorld, Visual Task Vectors, and Scaling Laws for Data Filtering.

Highlights of the latest release.

Highlights include Mixture of Depths, ReFT, and Many-shot Jailbreaking.

Highlights include Sora-artist collaboration, AIOS, BioMedLM, Mini-Gemini, sDPO, and Grok-1.5.

Highlights include FeatUp, EnvGen, RAFT, SceneScript, and Grok-1.

By combining Determined AI and Pachyderm open source community efforts we are enhancing the end-to-end AI workflow and driving greater productivity and success for our users.

Highlights of the latest release.