Tag: autoregressive models
Browse our exclusive articles!
Streaming Target Speaker Extraction with Chunk-wise Splicing
Discover a novel chunk-wise interleaved splicing method for real-time target speaker extraction using autoregressive models with high stability and low lat...
Reward Weighted Guidance for Efficient Autoregressive Models
Optimize autoregressive models with Reward Weighted Classifier-Free Guidance for faster policy improvement and flexible reward adaptation without retrainin...
Efficient Test-Time Search with Ordered Tokens in AR Models
Discover how coarse-to-fine ordered tokens improve test-time search efficiency and scalability in autoregressive generative models for image synthesis.
Introspective Diffusion Language Models: High-Quality NLP
Discover Introspective Diffusion Language Models (I-DLM) that combine diffusion and autoregressive strengths for superior NLP quality and efficiency.
WAND: Efficient Windowed Attention for Text-to-Speech
Discover WAND, a novel method using windowed attention and knowledge distillation to boost efficiency in autoregressive text-to-speech models.