Master LLMs with our FREE course in collaboration with Activeloop & Intel Disruptor Initiative. Join now!


Unlimiformer: Long-Range Transformers with Unlimited Length Input
Latest   Machine Learning

Unlimiformer: Long-Range Transformers with Unlimited Length Input

Last Updated on June 8, 2023 by Editorial Team

Author(s): Reza Yazdanfar

Originally published on Towards AI.

Now it’s possible to have deep learning models with no limitation for the input size.


Attention-based transformers have revolutionized the AI industry since 2017. Since then, we have seen significant progress in all aspects, including Computer vision, NLP, …

Attentions are considered a more powerful and capable version of Neural Networks for generalization on big datasets and are nothing more than routing between keys (K) and queries (Q), then non-linearity (Softmax), and then values (V).

Numerous variations of attention have been developed for various reasons, including optimizing complexity, decreasing computation, etc. One main branch is about transformers' capacity. How much can we give a model? How much can the model remember? …

My previous article was bout extending… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI

Feedback ↓