Aurora 0.7b.2 Download -

Because of its small size, the model runs within a browser using WebLLM, enabling interactive coding tutors.

Recent benchmarks on the GSM8K (math reasoning) and HellaSwag (commonsense inference) datasets show that Aurora 0.7b.2 scores within 10% of models ten times its size, thanks to its dense training regime. Aurora 0.7b.2 Download

Indie game developers are using the Aurora 0.7b.2 download to power real-time, dynamic non-player character conversations without GPU requirements. Because of its small size, the model runs

For users running the model locally via llama.cpp or text-generation-webui, quantized versions are preferred. Because of its small size