Merge branch 'upstream' into concedo_experimental

# Conflicts:
#	Makefile
#	README.md
#	common/CMakeLists.txt
#	common/common.cpp
#	common/common.h
#	examples/embedding/embedding.cpp
#	examples/imatrix/imatrix.cpp
#	examples/infill/infill.cpp
#	examples/parallel/parallel.cpp
#	examples/perplexity/perplexity.cpp
#	examples/rpc/README.md
#	examples/save-load-state/save-load-state.cpp
#	examples/server/README.md
#	examples/speculative/speculative.cpp
#	tests/test-sampling.cpp

examples/simple/simple.cpp

@@ -1,3 +1,4 @@
+#include "arg.h"
 #include "common.h"
 #include "llama.h"
 
@@ -18,8 +19,7 @@ int main(int argc, char ** argv) {
     params.prompt = "Hello my name is";
     params.n_predict = 32;
 
-    auto options = gpt_params_parser_init(params, LLAMA_EXAMPLE_COMMON, print_usage);
-    if (!gpt_params_parse(argc, argv, params, options)) {
+    if (!gpt_params_parse(argc, argv, params, LLAMA_EXAMPLE_COMMON, print_usage)) {
         return 1;
     }
 
@@ -118,8 +118,6 @@ int main(int argc, char ** argv) {
         {
             const llama_token new_token_id = llama_sampler_sample(smpl, ctx, batch.n_tokens - 1);
 
-            llama_sampler_accept(smpl, new_token_id);
-
             // is it an end of generation?
             if (llama_token_is_eog(model, new_token_id) || n_cur == n_predict) {
                 LOG_TEE("\n");