Georgi Gerganov
254098a279
* common : refactor common_sampler + grammar logic changes * tests : increase max_tokens to get needed response * batched : fix uninitialized samplers
llama.cpp/examples/lookup
Demonstration of Prompt Lookup Decoding
https://github.com/apoorvumang/prompt-lookup-decoding
The key parameters for lookup decoding are ngram_min, ngram_max and n_draft. The first two determine the size of the ngrams to search for in the prompt for a match. The latter specifies how many subsequent tokens to draft if a match is found.
More info:
https://github.com/ggml-org/llama.cpp/pull/4484 https://github.com/ggml-org/llama.cpp/issues/4226