CMU’s Unlimiformer Augments Transformers to Enable Unbounded Input Lengths

In the new paper Unlimiformer: Long-Range Transformers With Unlimited Length Input, a Carnegie Mellon University research team presents a general approach for improving model performance by augmenting pretrained encoder-decoder transformers with an external datastore to permit inputs of unbounded length.

ai artificial intelligence carnegie mellon carnegie mellon university decoder deep-neural-networks encoder encoder-decoder general machine learning machine learning & data science ml paper performance research research team team technology transformers university university research