DCKV learns end-to-end compression of hybrid-model prompt KV caches
On Qwen3.5-2B, tokens use over 55 times fewer bytes than bf16 with at most 1.8 points lost on three long-context benchmarks.
Stories per day · 30 days
Relatedautoregressive-models1chain-of-thought1compression1diffusion-models1kv-cache1long-context1multimodal-reasoning1qwen1
2 stories, newest first · one card per story
Stories tagged #hybrid-models
CoRAL pairs a diffusion planner with an autoregressive executor
A diffusion planner drafts K parallel reasoning paths; a large autoregressive model selects one and writes the answer.