Meta, the parent company of Facebook, has recently concluded the first-ever LlamaCon, a virtual conference that brings together developers worldwide. The program saw many notable Meta personalities debrief the audience, including Meta CEO and founder Mark Zuckerberg. Meta also took this opportunity to reveal new tools that can help developers build with the Llama LLM (Large Language Model). Here are some of the highlights from the event:
Meta has released a preview of the Llama API for free. The Llama API provides users a more convenient platform to use the various Llama models, including their latest models Llama Maverick and Llama Scout. The API has an SDK for both Python and Typescript and is compatible with the OpenAI SDK.
See Related: Meta Announces “Next Generation” AI Chip A Day After Intel And Google
Inference Speed And Partnership
Meta is also offering faster inference speed by partnering with Cerberus and Groq. Developers can select which model they want to use in the API. This feature is available by request. “This collaboration reinforces our commitment to fostering a diverse ecosystem that offers flexibility and choice”, the company stated.
There is also an expanded Llama Stack integration with NVIDIA and NeMo microservices. Additionally, Meta is working with other companies that will be included in the Stack in the future.
Meta also used the occasion to announce 10 international recipients of the Llama Impact Grants. The total grant money this year was over $1.5 million and was awarded to various start-ups, universities, and companies that have used Llama to create original and transformative AI solutions.
