From 689d377916947c303f6393629238727d0fb3753c Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Fri, 19 Jul 2024 14:21:33 +0300 Subject: [PATCH] cont ggml-ci --- examples/gbnf-validator/gbnf-validator.cpp | 4 ++-- src/llama-impl.h | 8 +------- tests/test-grammar-integration.cpp | 4 ++-- 3 files changed, 5 insertions(+), 11 deletions(-) diff --git a/examples/gbnf-validator/gbnf-validator.cpp b/examples/gbnf-validator/gbnf-validator.cpp index a52df566d..48a705e15 100644 --- a/examples/gbnf-validator/gbnf-validator.cpp +++ b/examples/gbnf-validator/gbnf-validator.cpp @@ -16,14 +16,14 @@ static bool llama_sample_grammar_string(struct llama_grammar * grammar, const st auto decoded = decode_utf8(input_str, {}); const auto & code_points = decoded.first; - const llama_grammar_rules & prev_rules = llama_grammar_get_rules (grammar); + const llama_grammar_rules & rules = llama_grammar_get_rules (grammar); llama_grammar_stacks & cur_stacks = llama_grammar_get_stacks(grammar); size_t pos = 0; for (auto it = code_points.begin(), end = code_points.end() - 1; it != end; ++it) { const llama_grammar_stacks prev_stacks = llama_grammar_get_stacks(grammar); // copy - llama_grammar_accept(prev_rules, prev_stacks, *it, cur_stacks); + llama_grammar_accept(rules, prev_stacks, *it, cur_stacks); if (cur_stacks.empty()) { error_pos = pos; diff --git a/src/llama-impl.h b/src/llama-impl.h index 2a9110e54..458049a97 100644 --- a/src/llama-impl.h +++ b/src/llama-impl.h @@ -3,12 +3,6 @@ #define LLAMA_API_INTERNAL #include "llama.h" -#include -#include -#include -#include -#include - #ifdef __has_include #if __has_include() #include @@ -24,7 +18,7 @@ // bump if necessary #define LLAMA_MAX_NODES 8192 -#define LLAMA_MAX_LAYERS 256 +#define LLAMA_MAX_LAYERS 512 #define LLAMA_MAX_EXPERTS 160 // DeepSeekV2 #ifdef __GNUC__ diff --git a/tests/test-grammar-integration.cpp b/tests/test-grammar-integration.cpp index 41ff0417f..68f971bfe 100644 --- a/tests/test-grammar-integration.cpp +++ b/tests/test-grammar-integration.cpp @@ -49,13 +49,13 @@ static bool match_string(const std::string & input, llama_grammar * grammar) { const auto & code_points = decoded.first; - const llama_grammar_rules & prev_rules = llama_grammar_get_rules (grammar); + const llama_grammar_rules & rules = llama_grammar_get_rules (grammar); llama_grammar_stacks & cur_stacks = llama_grammar_get_stacks(grammar); for (auto it = code_points.begin(), end = code_points.end() - 1; it != end; ++it) { const llama_grammar_stacks prev_stacks = llama_grammar_get_stacks(grammar); // copy - llama_grammar_accept(prev_rules, prev_stacks, *it, cur_stacks); + llama_grammar_accept(rules, prev_stacks, *it, cur_stacks); if (cur_stacks.empty()) { // no stacks means that the grammar failed to match at this point