StreamFlow achieves streaming speech token decoding with 180 ms first-packet latency by using hierarchical block-wise attention masks in a DiT flow matching model.
Training Setup For training, we use the Chinese and English subsets of Emilia [18], which totals 100,000 hours of speech data
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
StreamFlow: Streaming Flow Matching with Block-wise Guided Attention Mask for Speech Token Decoding
StreamFlow achieves streaming speech token decoding with 180 ms first-packet latency by using hierarchical block-wise attention masks in a DiT flow matching model.