Roman Potemin's picture

Roman Potemin

poteminr

·

poteminr

AI & ML interests

nlp, ner, multi-agent systems, code agents

Recent Activity

liked a model 10 days ago

facebook/dinov3-vitb16-pretrain-lvd1689m

liked a model 11 days ago

google/siglip2-base-patch16-224

upvoted a collection 12 days ago

View all activity

Organizations

None yet

upvoted a collection 12 days ago

DINOv3

DINOv3: foundation models producing excellent dense features, outperforming SotA w/o fine-tuning - https://arxiv.org/abs/2508.10104 • 15 items • Updated Mar 10 • 617

upvoted 2 collections 2 months ago

FastVLM

Efficient Vision Encoding for Vision Language Models • 8 items • Updated Mar 2 • 112

MobileCLIP2

MobileCLIP2: Mobile-friendly image-text models with SOTA zero-shot capabilities trained on DFNDR-2B • 30 items • Updated 6 days ago • 61

upvoted an article 2 months ago

Article

SigLIP 2: A better multilingual vision language encoder

+1

Feb 21, 2025

•

210

upvoted a paper 2 months ago

SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features

Paper • 2502.14786 • Published Feb 20, 2025 • 164

upvoted a collection 2 months ago

SigLIP2

36 items • Updated Mar 12 • 114

upvoted a collection about 1 year ago

ModernBERT

Bringing BERT into modernity via both architecture changes and scaling • 3 items • Updated Dec 19, 2024 • 158

upvoted a collection about 2 years ago

AQLM

AQLM quantized LLMs • 21 items • Updated Feb 18 • 46