Inquiry

Research

Research

Publications

Incoming

Research

Articles

Independent work that wasnt published, but has valuable insight.

Bar chart comparing LLaVA-1.5-7B inference latency for sequential and late-insertion approaches

Vision-Language Models · Draft

Speeding up VLM inference with late token insertion

An exploration of delaying selected visual tokens until they are needed, reducing wasted computation while preserving the model’s useful context.

Chart showing 96.4 percent full coverage when predicting the top 24 experts in a mixture-of-experts model

Mixture-of-Experts

Reducing VRAM footprint of MoE LLMs through predictive loading

A study of predicting which experts will be activated so model weights can be staged just ahead of demand instead of remaining resident in VRAM.