6 accepted papers
Diffusion Transformer models have driven the rapid advances in video generation, achieving state-of-the-art quality and flexibility. However, their attention mechanism remains a major performance bottleneck, as its dense computation scales quadratically with the sequence length. To overcome this lim…
Diffusion large language models (dLLMs) have emerged as a promising alternative for text generation, distinguished by their native support for parallel decoding. In practice, block inference is crucial for avoiding order misalignment in global bidirectional decoding and improving output quality. How…
Batch inference for agentic workloads stresses the GPU key–value (KV) cache in a sustained and cumulative manner, often causing severe throughput degradation well before memory capacity is exhausted. We identify this phenomenon as middle-phase thrashing, a previously under-characterized pathology in…
Speculative decoding mitigates the memory-bound nature of LLM decoding by using a lightweight draft model to propose multiple tokens for parallel verification. However, its adoption has been limited by the lack of high-quality draft models and scalable training infrastructure. We introduce SpecForge…
Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs. In this study, we introduce GliDe and CaPE, two low-hassle modifications to vanilla speculative decoding to further improve the decoding speed of a frozen LLM. S…
Transformer achieves promising results on various tasks. However, self-attention suffers from quadratic memory requirements with respect to the sequence length. Existing work focuses on reducing time and space complexity from an algorithm perspective. In this work, we propose sequence parallelism, a…