#speculative-decoding - Tags

4 posts · Transformer Series

Tag: #speculative-decoding

🗓 2026-04-11 • Transformer Series • ⏱ 63 min read

Speculative decoding, speculative sampling, blockwise parallel decoding, token tree verification, and Hugging Face implementation details.

🗓 2026-04-11 • Transformer Series • ⏱ 40 min read

DeepSeek MTP: EAGLE, HASS, classical multi-token prediction, DeepSeek’s causal-chain design, formulas, and the vLLM implementation.

🗓 2026-04-11 • Transformer Series • ⏱ 29 min read

Lookahead decoding: Jacobi decoding, n-gram pool, 2D window, parallel verification, and llama.cpp implementation details.

🗓 2026-04-11 • Transformer Series • ⏱ 32 min read

Medusa: multi-decoding heads, tree attention, typical acceptance, sparse tree construction, training strategies, and decoding flow.