As described in
Longformer: The Long-Document Transformer by Iz Beltagy, Matthew E. Peters, Arman Cohan,
led-large-16384 was initialized from
bart-large since both models share the exact same architecture. To be able to process 16K tokens,
bart-large's position embedding matrix was simply copied 16 times.
This model is especially interesting for long-range summarization and question answering.
This notebook shows how
led-large-16384 can effectively be fine-tuned on a downstream task.
