diff options
author | Justin Suess <justin.suess@westpoint.edu> | 2023-03-15 15:42:40 -0400 |
---|---|---|
committer | GitHub <noreply@github.com> | 2023-03-15 21:42:40 +0200 |
commit | 2d64715ad475f192a4004a52d134c67ccb6f44ad (patch) | |
tree | 1809c957a8cd0b7fd530e18288ed2d925041218a /main.cpp | |
parent | 16b2c61a22f828ea77d9f084ca871c63bc5cc283 (diff) |
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>
Diffstat (limited to 'main.cpp')
-rw-r--r-- | main.cpp | 5 |
1 files changed, 3 insertions, 2 deletions
@@ -547,6 +547,8 @@ bool llama_eval( const int d_key = n_embd/n_head; + // TODO: check if this size scales with n_ctx linearly and remove constant. somehow I feel it wasn't the case + // static size_t buf_size = hparams.n_ctx*1024*1024; static size_t buf_size = 512u*1024*1024; static void * buf = malloc(buf_size); @@ -819,8 +821,7 @@ int main(int argc, char ** argv) { // load the model { const int64_t t_start_us = ggml_time_us(); - - if (!llama_model_load(params.model, model, vocab, 512)) { // TODO: set context from user input ?? + if (!llama_model_load(params.model, model, vocab, params.n_ctx)) { fprintf(stderr, "%s: failed to load model from '%s'\n", __func__, params.model.c_str()); return 1; } |