A CTC-based draft model with sequence-level training and a CTC transform for candidate cleaning improves accepted-token counts and inference speedup in speculative decoding of LLMs.
Tutorial on directed acyclic graphs
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Speculative Decoding with CTC-based Draft Model for LLM Inference Acceleration
A CTC-based draft model with sequence-level training and a CTC transform for candidate cleaning improves accepted-token counts and inference speedup in speculative decoding of LLMs.