Meta may be releasing a smaller Llama 3
Briefly

Meta is preparing to release smaller versions of its Llama language model to meet the increasing demand for cost-effective AI models. These models are faster, more flexible, and cheaper to run than regular-sized models.
Smaller AI models are powerful, capable of tasks like summarizing PDFs, conversations, and writing code. These lightweight models are suitable for specific projects and devices that require less computing power.
Lightweight AI models are beneficial for users who do not need the extensive capabilities of larger models. They can be deployed in areas like code assistance and on devices with limitations in power usage, such as phones and laptops.
Meta intends to roll out Llama 3, a smaller and potentially more open model than Llama 2, in July. This version may address controversial questions that the previous iteration was restricted from answering.
Read at The Verge
[
|
]