Be a part of our each day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Study Extra
Microsoft isn’t resting its AI success on the laurels of its partnership with OpenAI.
No, removed from it. As an alternative, the corporate typically referred to as Redmond for its headquarters location in Washington state as we speak got here out swinging with the discharge of three new fashions in its evolving Phi collection of language/multimodal AI.
The three new Phi 3.5 fashions embody the three.82 billion parameter Phi-3.5-mini-instruct, the 41.9 billion parameter Phi-3.5-MoE-instruct, and the 4.15 billion parameter Phi-3.5-vision-instruct, every designed for primary/quick reasoning, extra highly effective reasoning, and imaginative and prescient (picture and video evaluation) duties, respectively.
All three fashions can be found for builders to obtain, use, and fine-tune customise on Hugging Face underneath a Microsoft-branded MIT License that enables for industrial utilization and modification with out restrictions.
Amazingly, all three fashions additionally boast close to state-of-the-art efficiency throughout various third-party benchmark exams, even beating different AI suppliers together with Google’s Gemini 1.5 Flash, Meta’s Llama 3.1, and even OpenAI’s GPT-4o in some instances.
That efficiency, mixed with the permissive open license, has folks praising Microsoft on the social community X:
Let’s overview every of the brand new fashions as we speak, briefly, primarily based on their launch notes posted to Hugging Face
Phi-3.5 Mini Instruct: Optimized for Compute-Constrained Environments
The Phi-3.5 Mini Instruct mannequin is a light-weight AI mannequin with 3.8 billion parameters, engineered for instruction adherence and supporting a 128k token context size.
This mannequin is right for eventualities that demand robust reasoning capabilities in memory- or compute-constrained environments, together with duties like code technology, mathematical drawback fixing, and logic-based reasoning.
Regardless of its compact dimension, the Phi-3.5 Mini Instruct mannequin demonstrates aggressive efficiency in multilingual and multi-turn conversational duties, reflecting vital enhancements from its predecessors.
It boasts near-state-of-the-art efficiency on various benchmarks and overtakes different similarly-sized fashions (Llama-3.1-8B-instruct and Mistral-7B-instruct) on the RepoQA benchmark which measures “long context code understanding.”
Phi-3.5 MoE: Microsoft’s ‘Mixture of Experts’
The Phi-3.5 MoE (Combination of Consultants) mannequin seems to be the primary on this mannequin class from the agency, one that mixes a number of completely different mannequin varieties into one, every specializing in numerous duties.
This mannequin leverages an structure with 42 billion energetic parameters and helps a 128k token context size, offering scalable AI efficiency for demanding purposes. Nonetheless, it operates nly with 6.6B energetic parameters, in response to the HuggingFace documentation.
Designed to excel in numerous reasoning duties, Phi-3.5 MoE affords robust efficiency in code, math, and multilingual language understanding, typically outperforming bigger fashions in particular benchmarks, together with, once more, RepoQA:
It additionally impressively beats GPT-4o mini on the 5-shot MMLU (Large Multitask Language Understanding) throughout topics reminiscent of STEM, the humanities, the social sciences, at various ranges of experience.
The MoE mannequin’s distinctive structure permits it to keep up effectivity whereas dealing with complicated AI duties throughout a number of languages.
Phi-3.5 Imaginative and prescient Instruct: Superior Multimodal Reasoning
Finishing the trio is the Phi-3.5 Imaginative and prescient Instruct mannequin, which integrates each textual content and picture processing capabilities.
This multimodal mannequin is especially suited to duties reminiscent of normal picture understanding, optical character recognition, chart and desk comprehension, and video summarization.
Like the opposite fashions within the Phi-3.5 collection, Imaginative and prescient Instruct helps a 128k token context size, enabling it to handle complicated, multi-frame visible duties.
Microsoft highlights that this mannequin was skilled with a mix of artificial and filtered publicly accessible datasets, specializing in high-quality, reasoning-dense information.
Coaching the brand new Phi trio
The Phi-3.5 Mini Instruct mannequin was skilled on 3.4 trillion tokens utilizing 512 H100-80G GPUs over 10 days, whereas the Imaginative and prescient Instruct mannequin was skilled on 500 billion tokens utilizing 256 A100-80G GPUs over 6 days.
The Phi-3.5 MoE mannequin, which includes a mixture-of-experts structure, was skilled on 4.9 trillion tokens with 512 H100-80G GPUs over 23 days.
Open-source underneath MIT License
All three Phi-3.5 fashions can be found underneath the MIT license, reflecting Microsoft’s dedication to supporting the open-source group.
This license permits builders to freely use, modify, merge, publish, distribute, sublicense, or promote copies of the software program.
The license additionally features a disclaimer that the software program is supplied “as is,” with out warranties of any variety. Microsoft and different copyright holders are usually not accountable for any claims, damages, or different liabilities that will come up from the software program’s use.
Microsoft’s launch of the Phi-3.5 collection represents a major step ahead within the growth of multilingual and multimodal AI.
By providing these fashions underneath an open-source license, Microsoft empowers builders to combine cutting-edge AI capabilities into their purposes, fostering innovation throughout each industrial and analysis domains.
