| so the use of llms in daily life is increasing however only english is the language of all the base models , plus also the current state of the art llms | |
| are created using the transformer architecture, which has proven to be the industry standard however due to its self attention mechanism | |
| its been computationally inefficient so we are proposing Cauvery 7b , a 7 billion parameter large language model currently under development | |
| that DOES NOT USE THE TRANSFORMER ARCHITECTURE AND AS AN ALTERNATIVE USES THE retentive network architecture with retention mechanism, we are | |
| in our early stages and looking for investors |