Mistral 7B Instruct v0.2
Readymodel path:accounts/fireworks/models/mistral-7b-instruct-v0p2
Mistral 7B Instruct v0.2 is an instruction fine-tuned version of the Mistral 7B v0.2 language model. This update introduces key improvements over v0.1, including an expanded context window of 32k tokens (up from 8k), a Rope-theta value of 1e6, and the removal of Sliding-Window Attention. Designed to excel in generating coherent and contextually rich responses, it is suitable for a wide range of natural language processing tasks.
On-demand deployments allow you to use Mistral 7B Instruct v0.2 on dedicated GPUs with Fireworks' high-performance serving stack with high reliability and no rate limits.