Skip to main content

Speculative decoding

Inferlets that implement and verify draft tokens before committing them.

Running on the dummy driver?

The dummy driver samples a fresh random token for every slot, so every draft is rejected and these inferlets run in 1-token-per-step fallback — they complete correctly but no speedup is observable.

InferletWhat it shows
cacheback-speculative-decodingPrompt-lookup n-gram drafting with batched target verification.
mtp-speculative-decodingNative MTP drafting with explicit accept, correction, and bonus-token handling.