dqnGPT-v1 is the human-like, general assistant of the DQN suite. Based on the new Gemma 4 architecture, this 8B E4B model is excellent for daily use.
AI & ML interests
We do LLMs and reinforcement learning! Everybody welcome to contribute!
Recent Activity
View all activity
dqnGPT-v1 is the human-like, general assistant of the DQN suite. Based on the new Gemma 4 architecture, this 8B E4B model is excellent for daily use.
dqnCode-v1 is a bespoke 4B coding focused model. With a score of 63% on HumanEval, dqnCode has top-tier coding performance for local AI deployment.
dqnMath-v1 is an efficient 4B math model, trained for token-efficient math solving, targeted at daily-use, grade school math, for local deployment.