SPOT predicts commit order to accelerate diffusion language decoding
On SDAR-8B, SPOT cuts target passes by up to 6.8× and reaches up to 3.4× standard throughput.
7 stories, newest first · one card per story
On SDAR-8B, SPOT cuts target passes by up to 6.8× and reaches up to 3.4× standard throughput.
DCOCO restores intra-block context in parallel speculative drafts, reaching up to 6.6x decoding speedup.
TokenRouter raises token-level LLM routing decode throughput by 2.01–64.15× over existing serving systems.
Mean speedups versus autoregressive decoding are 3.84 times on Qwen3-4B and 4.14 times on Qwen3-8B.
TRIAGE stabilizes native NVFP4 RL on Qwen3 models, matching full precision with up to 2.3x rollout throughput.
AgSpec speeds coding-agent generation via retrieval and adaptive speculative drafts, up to 4.76× throughput.
PEARL adapts prefill-decode execution and elastic GPUs, raising agentic RL throughput by up to 2.79x.
Today 2.6K readers · 2.8K views · 145 reads · all time 206K views