Tae-Ho Kim

Nota AI

|

CTO & Co-Founder

Insight

Two papers discussing quantization methods specialized for MoE LLMs, accepted to an ICML 2026 workshop

min read

Two papers discussing quantization methods specialized for MoE LLMs, accepted to an ICML 2026 workshop

ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models

February 26, 2026

min read

ERGO: Efficient High-Resolution Visual Understanding for Vision-Language Models

NotaMoEQuantization: An MoE-Specific Quantization Method for Solar-Open-100B

min read

NotaMoEQuantization: An MoE-Specific Quantization Method for Solar-Open-100B

UniForm: A Reuse Attention Mechanism for Efficient Transformers on Resource-Constrained Edge Devices

min read

UniForm: A Reuse Attention Mechanism for Efficient Transformers on Resource-Constrained Edge Devices

Video Self-Distillation for Single-Image Encoders: Learning Temporal Priors from Unlabeled Video

min read

Video Self-Distillation for Single-Image Encoders: Learning Temporal Priors from Unlabeled Video

Efficient LLaMA-3.2-Vision by Trimming Cross-attended Visual Features

min read

Efficient LLaMA-3.2-Vision by Trimming Cross-attended Visual Features

Culture

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.