Unum develops in-memory vector search technology for navigating datasets with high-dimensional vectors, enabling real-time semantic search for large-scale marketplaces and content providers.

Unum

Unum is a deep-tech company that designs core software technologies for next-generation Data-Lakes and Cloud providers.

Unum's portfolio of products includes hardware-accelerated Storage Engines, Compute Kernels, Search Algorithms, and Neural Networks that can process Google-scale datasets in a fraction of the time and at lower costs.

Today, we are releasing a new set of pocket-sized multimodal AI models, trained in partnership with Nebius and already available on Hugging Face 🤗

- Matryoshka style multimodal embeddings ranging from 64 to 256 and 768 dimensions 🖼️

- Improved multimodal chat in 1.2B parameters, aligned with Direct Preference Optimization 💬

- ONNX backend, making PyTorch dependency optional for lightning-fast deployments ⚡

This marks our biggest AI release to date, paving the way for real-time multimodal perception and personalized assistants that can run on any device, including Windows, Linux, macOS, iOS, Android, and most wearable and IoT devices.

Tuning with Direct Preference Optimization allowed us to grow our "Multi-Modal Evaluation" (MME) perception score from 863 to 1049 for the same baseline model. Unexpectedly significant for us and visible to the naked eye. Avoiding PyTorch dependency allowed us to shrink the image size from over 5 GB to under 500 MB and will shrink further. Training with Matryoshka losses allows you to crop resulting embeddings, dropping up to 92% of the resulting dimensions but often retaining 99% of the original accuracy. 

The new guide also covers the recommended quantization and down-casting approaches, helping you export `f32`, `f16`, `i8`, and binary vectors and use them in conjunction with USearch and the rest of our hardware-friendly ecosystem of Search & AI infra 🤗

New models: https://lnkd.in/dpfGcV59
GitHub repository: https://lnkd.in/dTrZ5Q2d

How to make language models more compact using Nebius AI?⚡️

Our colleagues at Unum just did exactly that on our infrastructure. The goal was to scale down UForm-Gen, which is already among the industry's smallest multimodal generative models. The smaller it becomes, the easier it is to deploy on edge and mobile devices, popularizing local, privacy-preserving AI.

UNUM pre-trained the model on an internal image captioning dataset, trained on H100, and then fine-tuned it on public instruction datasets such as SVIT, LVIS, and VQAs. 

The resulting UForm-Gen-2 has just been released on Hugging Face: https://lnkd.in/dyUjQVdF. It works with the HF Transformers library out of the box, free for commercial use. The model also significantly outperforms its predecessor in visual question answering and multi-turn chat capabilities.

If you are facing the task of training ML models, learn how to do it effectively with Nebius AI: https://lnkd.in/dezbZsPQ

How small should a Language Model be?

Less than two months ago, we released UForm-Gen, one of the industry's most miniature multimodal Generative #AI models. Now downloaded over 100,000 times a month, it's one of the industry's most popular captioning and visual question-answering models. It was built on a Language Model with only 1.3 billion parameters in size and sometimes worked better for Vision-Language tasks than the 100-1000x larger Google Gemini.

Scaling down is the key to privacy-preserving AI models running on every one of the billions of chips produced yearly! So, using the NVIDIA DGX-H100 nodes on the Nebius #GPU cloud, we've trained several tiny models! One of them features the smallest Language Model we've used to date, derived from the "Qwen1.5" by Alibaba Group with only 0.5 billion parameters.

The model significantly outperforms our previous result thanks to a considerably larger and higher-quality multimodal dataset and an improved Vision tower!

On a technical side, the new "UForm-Gen-2" scored 19.58 on MM-Vet, 45.5 on SQA, and 880 on MME benchmarks, similar to last year's models 10-20x the size. It works with the Hugging Face Transformers library out of the box and is already available online, free for commercial use 🤗

https://lnkd.in/gti4HdAK

UForm is going Generative!

The UForm family of tiny multimodal AI models just got broader! In addition to the existing CLIP-like embedding models, we now have a generative model useful for image captioning, visual question answering, and multimodal chats. All that is #opensource and takes around a billion parameters, small enough to fit even on mobile devices 🎉

Repository: https://lnkd.in/dTrZ5Q2d
Generative model: https://lnkd.in/gZ9y4KEW
Chat model: https://lnkd.in/gpaRVvKm
Discord: https://lnkd.in/gGj-rRGW

Check our the quality of image captions in the comments ⬇️

Billion Scale Vector Search for #AI. 10x Faster than Meta's FAISS, 1000x Faster than Pinecone.

Thanks to our Intel Corporation and Gcore  partners for providing the hardware and help with the benchmarks.

https://lnkd.in/gQK9TViT

news_impact_assignment_basic

key_people

merger_acquisitions

network_dossier

public_peers

news_daily_refresh

news_processing_daily_refresh

news_impact_assignment_daily

news_processing_basic

legal

legal_dossier

bundesanzeiger

events_dossier

partnerships_market_presence_dossier

partnerships_market_presence

company_dossier

network

strategic_challenges

industry_trends

strategic_challenges_dossier

industry_trends_dossier

company_summary

events

network_mapping

transactions_dossier

swot_dossier

signals

financial_statements

crunchbase_trigger

crunchbase_profile

li_profile

financials_download_nd

private_financials

swot_analysis

financial_legal_advisories

company_structure

industry_analysis

financial_analysis

funding_rounds_analysis

regulatory_analysis

competition_analysis

news_impact_assignment_advanced

news_processing_advanced

news_advanced

purchaser_profile

network_onboarding

translate_crunchbase_fundingrounds

peers

customer_profile

news_basic

valuation

news impact assignment basic	general
public peers	general	Sep 2, 2024 11:39 AM	3d98a0fe
news daily refresh	general
news processing daily refresh	general
news impact assignment daily	general
news processing basic	general	Sep 3, 2024 8:18 AM	0dc20e7c
company summary	general
signals	general
crunchbase trigger	general	Nov 27, 2024 12:50 PM	e8a1fd51
crunchbase profile	general
li profile	general	Sep 2, 2024 11:39 AM	2e1c2c06
translate crunchbase fundingrounds	general
peers	general	Sep 2, 2024 11:39 AM	d4cf349d
news basic	general	Sep 2, 2024 11:39 AM	38ab5ffa
bundesanzeiger	filings
financials download nd	financials
private financials	financials

CLOUD:UNUM

Unum