About
We created an LLM that outperform OpenELM and Phi on MT-Bench, in just 9 days. It's built on the Lightning framework with optimisations from TinyLlama, achieving ultra high throughput (~99.6% GPU util
About
We created an LLM that outperform OpenELM and Phi on MT-Bench, in just 9 days. It's built on the Lightning framework with optimisations from TinyLlama, achieving ultra high throughput (~99.6% GPU util