ALiBi enables transformers trained on length-1024 sequences to extrapolate to length-2048 with the same perplexity as a sinusoidal model trained on 2048, while training 11% faster and using 11% less memory.
CAPE: encoding relative positions with continuous augmented positional embeddings
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2021 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation
ALiBi enables transformers trained on length-1024 sequences to extrapolate to length-2048 with the same perplexity as a sinusoidal model trained on 2048, while training 11% faster and using 11% less memory.