llama-cli : fix dangling reference to sampler config

Author: ggerganov

common/common.cpp

@@ -976,8 +976,6 @@ struct common_init_result common_init_context_from_model(
     const llama_vocab * vocab = llama_model_get_vocab(model);
 
     auto cparams = common_context_params_to_llama(params);
-    cparams.samplers = params.backend_samplers;
-    cparams.n_samplers = params.n_backend_samplers;
 
     llama_context * lctx = llama_init_from_model(model, cparams);
     if (lctx == NULL) {
@@ -1247,6 +1245,9 @@ struct llama_context_params common_context_params_to_llama(const common_params &
     cparams.type_k = params.cache_type_k;
     cparams.type_v = params.cache_type_v;
 
+    cparams.samplers   = params.backend_samplers;
+    cparams.n_samplers = params.n_backend_samplers;
+
     return cparams;
 }
 

tools/main/main.cpp

@@ -146,8 +146,9 @@ int main(int argc, char ** argv) {
 
     // Configure backend sampler if configured
     llama_sampler * backend_sampler = common_sampler_backend_init(model, sparams);
+    llama_sampler_seq_config sampler_config = { 0, backend_sampler };
+
     if (backend_sampler) {
-        llama_sampler_seq_config sampler_config = { 0, backend_sampler };
         params.backend_samplers = &sampler_config;
         params.n_backend_samplers = 1;
     }