From 0a243da7d470be6df312fdbe0c0fedd536c2c347 Mon Sep 17 00:00:00 2001 From: Minsoo Cheong Date: Sat, 23 Mar 2024 00:24:01 +0900 Subject: [PATCH] fixes based on review @cebtenzzre --- common/sampling.cpp | 8 ++++---- common/sampling.h | 2 +- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/common/sampling.cpp b/common/sampling.cpp index c9285892e..516868079 100644 --- a/common/sampling.cpp +++ b/common/sampling.cpp @@ -173,7 +173,7 @@ static llama_token llama_sampling_sample_impl( const float mirostat_tau = params.mirostat_tau; const float mirostat_eta = params.mirostat_eta; - std::vector original_logits = {}; + std::vector original_logits; auto cur_p = llama_sampling_configure_token_candidates(ctx_sampling, ctx_main, ctx_cfg, idx, !is_resampling, &original_logits); if (!is_resampling) { GGML_ASSERT(!original_logits.empty()); @@ -251,7 +251,7 @@ static llama_token_data_array llama_sampling_configure_token_candidates_impl( struct llama_context * ctx_cfg, const int idx, bool apply_grammar, - std::vector* original_logits) { + std::vector * original_logits) { const llama_sampling_params & params = ctx_sampling->params; const int n_vocab = llama_n_vocab(llama_get_model(ctx_main)); @@ -271,7 +271,7 @@ static llama_token_data_array llama_sampling_configure_token_candidates_impl( if (apply_grammar && original_logits != NULL) { // Only make a copy of the original logits if we are not applying grammar checks, not sure if I actually have to do this. - (*original_logits).insert((*original_logits).end(), logits, logits + llama_n_vocab(llama_get_model(ctx_main))); + *original_logits = {logits, logits + llama_n_vocab(llama_get_model(ctx_main))}; } // apply params.logit_bias map @@ -335,7 +335,7 @@ llama_token_data_array llama_sampling_configure_token_candidates( struct llama_context * ctx_cfg, const int idx, bool apply_grammar, - std::vector* original_logits) { + std::vector * original_logits) { return llama_sampling_configure_token_candidates_impl(ctx_sampling,ctx_main, ctx_cfg, idx, apply_grammar, original_logits); } diff --git a/common/sampling.h b/common/sampling.h index 38557a67d..89a6bdf91 100644 --- a/common/sampling.h +++ b/common/sampling.h @@ -138,7 +138,7 @@ llama_token_data_array llama_sampling_configure_token_candidates( struct llama_context * ctx_cfg, int idx = 0, bool apply_grammar = true, - std::vector* original_logits = nullptr); + std::vector * original_logits = nullptr); void llama_sampling_accept( struct llama_sampling_context * ctx_sampling,