d1-3B

Headlines · 2

1. 18h
   Notebooks released for fine-tuning the d1-3B decision model

   A post reposted by UnslothAI on X points to two notebooks in the Liquid4All cookbook for fine-tuning the d1-3B decision model on your own data, one built on Unsloth and one on Hugging Face TRL. The post says the notebooks cover preparing training data and the fine-tuning workflow, and claims the process fits in 30 minutes on a free-tier GPU. The time and ease claims come from the post and have not been independently verified.

   UnslothAI (X) · 🔥 6
2. Oct 7
   Liquid AI open-sources Open d1 decision models with zero output tokens

   Liquid AI released the open-weight Open d1 family of decision models: d1-3B (3.12B parameters, reads text and images) and d1-omni-600M (587M, an early research version that reads text with either images or audio). Rather than generating text, the models read a state once and return typed, calibrated probabilities for named questions, with single-question latency quoted as low as 8 ms, aimed at routing, moderation, intent classification, reranking, LLM-as-a-judge scoring and agent guardrails. Liquid AI says d1-3B scores 48.57 on Decision Index v0.2.1 using its own official scorer (not a leaderboard submission), above all sub-10B models; both checkpoints are on Hugging Face under the LFM Open License v1.0, free for commercial use below $10M annual revenue.

   Hugging Face · 🔥 0

Experience and discussion from the community

Share my d1-3B experienceAsk about d1-3B