minor fixes

common/common.cpp

@@ -1323,7 +1323,7 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
     printf("                        files containing context to embed.\n");
     printf("  --chunk-size N        minimum length of embedded text chunk (default:%d)\n", params.chunk_size);
     printf("  --chunk-separator STRING\n");
-    printf("                        string to separate chunks (default: newline)\n");
+    printf("                        string to separate chunks (default: \"\\n\")\n");
     printf("  -n N, --n-predict N   number of tokens to predict (default: %d, -1 = infinity, -2 = until context filled)\n", params.n_predict);
     printf("  -c N, --ctx-size N    size of the prompt context (default: %d, 0 = loaded from model)\n", params.n_ctx);
     printf("  -b N, --batch-size N  logical maximum batch size (default: %d)\n", params.n_batch);

examples/retrieval/retrieval.cpp

@@ -8,11 +8,11 @@ struct chunk {
     // filename
     std::string filename;
     // original file position
-    int64_t filepos;
+    size_t filepos;
     // original text data
     std::string textdata = "";
     // tokenized text data
-    std::vector<std::int32_t> tokens;
+    std::vector<llama_token> tokens;
     // embedding
     std::vector<float> embedding;
     // cosin similarity
@@ -240,9 +240,6 @@ int main(int argc, char ** argv) {
     while (true) {
         printf("Enter query: ");
         std::getline(std::cin, query);
-        if (query == "exit" || query == "quit" || query == "q") {
-            break;
-        }
         std::vector<int32_t> query_tokens = llama_tokenize(ctx, query, true);
 
         struct llama_batch query_batch = llama_batch_init(n_batch, 0, 1);