Story

arxiv_cs_cl ยท Nov 18, 2026 ยท paper

Source brief

LoopCache: Accelerating Long-Context Self-Speculative Decoding for Looped Transformers

arxiv.orgNov 18, 2026
original source linked

In brief

Looped Transformers achieve large effective depth with compact models by repeatedly applying a weight-shared block across R recurrent steps, each maintaining a loop-specific key-value (KV) cache. To accelerate sequent...

Continues in

Speculative Decoding โ€” open the evidence trace โ†’

Continue reading

Read the original at arxiv.org โ†’Open in live feedRead that dayโ€™s brief

Earlier in this thread 4 items