Microsoft's Phi-3 shows the surprising power of small, locally run AI language models
Briefly

from Ars Technica3 weeks ago
Microsoft introduces Phi-3-mini, a lightweight AI language model with 3.8 billion parameters, ideal for smartphones and laptops without internet connection.
Ars Technicahttps://arstechnica.com/information-technology/2024/04/microsofts-phi-3-shows-the-surprising-power-of-small-locally-run-ai-language-models/
Traditional large language models (LLMs) like Google's PaLM 2 and OpenAI's GPT-4 require heavy-duty data center GPUs, while Phi-3-mini is designed for consumer GPUs.
Ars Technicahttps://arstechnica.com/information-technology/2024/04/microsofts-phi-3-shows-the-surprising-power-of-small-locally-run-ai-language-models/
Microsoft's Phi-3-mini is part of the Phi series, following Phi-2 and Phi-1, and has a 4,000-token context window with plans for larger parameter versions.
Ars Technicahttps://arstechnica.com/information-technology/2024/04/microsofts-phi-3-shows-the-surprising-power-of-small-locally-run-ai-language-models/
Microsoft aims to release 7-billion and 14-billion parameter versions of Phi-3, claiming they will be more capable than the initial Phi-3-mini model.
Ars Technicahttps://arstechnica.com/information-technology/2024/04/microsofts-phi-3-shows-the-surprising-power-of-small-locally-run-ai-language-models/
Read at Ars Technica
#phi-3-mini #ai-language-models #parameter-count #microsoft #technology
[
add
]
[
|
|
]