fix perplexity after c-api refactor (#390)

* preallocate a buffer of fitting size for tokenization (utils.cpp)

* don't create a new std::string (especially here, where it's usually large)

1 files changed, 3 insertions, 1 deletions

diff --git a/utils.cpp b/utils.cpp
index 1679ae1..3909c97 100644
--- a/utils.cpp
+++ b/utils.cpp
@@ -146,8 +146,10 @@ std::string gpt_random_prompt(std::mt19937 & rng) {
 
 // TODO: not great allocating this every time
 std::vector<llama_token> llama_tokenize(struct llama_context * ctx, const std::string & text, bool add_bos) {
-    std::vector<llama_token> res(8096);
+    // initialize to prompt numer of chars, since n_tokens <= n_prompt_chars
+    std::vector<llama_token> res(text.size() + (int)add_bos);
     int n = llama_tokenize(ctx, text.c_str(), res.data(), res.size(), add_bos);
+    assert(n >= 0);
     res.resize(n);
 
     return res;