Routing early CLAP stages to the audio projector and later stages to the LLM yields encoder-free captioning that beats LLM-only distillation by +12.2 CIDEr-D on AudioCaps.
CED: consistent ensemble distillation for audio tagging,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
CARD: Cross-component Audio Representation Distillation for Encoder-Free Audio Captioning
Routing early CLAP stages to the audio projector and later stages to the LLM yields encoder-free captioning that beats LLM-only distillation by +12.2 CIDEr-D on AudioCaps.