Mistral's new speech model can run on a smartwatch or a smartphone.
Relatively light at just 2 billion parameters, the model is meant for use with consumer-grade GPUs for those who want to self ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results