DeepScaleR-1.5B-Preview
DeepScaleR-1.5B-Preview enhances math problem solving through distributed reinforcement learning, ideal for researchers developers and math competitors.
What is DeepScaleR-1.5B-Preview?
DeepScaleR-1.5B-Preview is an advanced large language model optimized for mathematical problem-solving through distributed reinforcement learning. It excels in long text reasoning and is tailored for educational applications, particularly in math education and competitions. This model, developed by the Sky Computing Lab and Berkeley AI Research, is available under the MIT license for researchers and developers.