Brief

Researchers have developed a new framework called "this http URL" designed to optimize the inference of diffusion large language models (dLLMs) on mobile devices. This framework addresses challenges such as shrinking workloads and complex data management inherent in mobile NPU architectures. It employs techniques like multi-block speculative decoding and dual-path progressive revision to significantly reduce generation latency while maintaining output quality. AI

Researchers have developed a new method called CreditDecoding to accelerate the text generation process in diffusion large language models (dLLMs). This technique addresses an inefficiency where models predict correct tokens earlier than their confidence scores allow for decoding, leading to redundant iterations. CreditDecoding quantifies a token's decoding potential using "Trace Credit" and fuses this with current model outputs to boost confidence in correct but underconfident tokens. This training-free approach has demonstrated significant speedups of up to 5.48 times with improved accuracy on various benchmarks and dLLM architectures. AI