added ctx_size parameter (#148)

* added ctx_size parameter

* added it in more places

* Apply suggestions from code review

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

1 files changed, 2 insertions, 1 deletions

diff --git a/utils.h b/utils.h
index 4f98011c..021120b0 100644
--- a/utils.h
+++ b/utils.h
@@ -17,7 +17,8 @@ struct gpt_params {
     int32_t n_threads = std::min(4, (int32_t) std::thread::hardware_concurrency());
     int32_t n_predict = 128; // new tokens to predict
     int32_t repeat_last_n = 64;  // last n tokens to penalize
-
+    int32_t n_ctx = 512; //context size
+    
     // sampling parameters
     int32_t top_k = 40;
     float   top_p = 0.95f;