From b634f8a26fef65210fd9fb2f87e83a2809535e89 Mon Sep 17 00:00:00 2001 From: Diego Devesa Date: Sat, 2 Nov 2024 13:08:53 +0100 Subject: [PATCH] simple-chat : only add bos on first prompt (#10129) --- examples/simple-chat/simple-chat.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/simple-chat/simple-chat.cpp b/examples/simple-chat/simple-chat.cpp index 14264cfcb..5f9973163 100644 --- a/examples/simple-chat/simple-chat.cpp +++ b/examples/simple-chat/simple-chat.cpp @@ -96,7 +96,7 @@ int main(int argc, char ** argv) { // tokenize the prompt const int n_prompt_tokens = -llama_tokenize(model, prompt.c_str(), prompt.size(), NULL, 0, true, true); std::vector prompt_tokens(n_prompt_tokens); - if (llama_tokenize(model, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), true, true) < 0) { + if (llama_tokenize(model, prompt.c_str(), prompt.size(), prompt_tokens.data(), prompt_tokens.size(), llama_get_kv_cache_used_cells(ctx) == 0, true) < 0) { GGML_ABORT("failed to tokenize the prompt\n"); }