wip

Makefile

@@ -770,6 +770,8 @@ main: tools/completion/completion.cpp common/arg.cpp common/download.cpp build-i
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 mainvk: tools/completion/completion.cpp common/arg.cpp common/download.cpp build-info.h ggml_v4_vulkan.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o console.o llavaclip_vulkan.o llava.o ggml-backend_vulkan.o ggml-backend-reg_vulkan.o ggml-vulkan.o ggml-vulkan-shaders.o ggml-repack.o $(OBJS_FULL) $(OBJS) lib/vulkan-1.lib
 	$(CXX) $(CXXFLAGS) -DGGML_USE_VULKAN -DSD_USE_VULKAN $(filter-out %.h,$^) -o $@ $(LDFLAGS)
+fitparams: tools/fit-params/fit-params.cpp common/arg.cpp common/download.cpp build-info.h ggml_v4_vulkan.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o console.o llavaclip_vulkan.o llava.o ggml-backend_vulkan.o ggml-backend-reg_vulkan.o ggml-vulkan.o ggml-vulkan-shaders.o ggml-repack.o $(OBJS_FULL) $(OBJS) lib/vulkan-1.lib
+	$(CXX) $(CXXFLAGS) -DGGML_USE_VULKAN -DSD_USE_VULKAN $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 sdmain: otherarch/sdcpp/util.cpp otherarch/sdcpp/main.cpp otherarch/sdcpp/stable-diffusion.cpp otherarch/sdcpp/upscaler.cpp otherarch/sdcpp/model.cpp otherarch/sdcpp/name_conversion.cpp otherarch/sdcpp/tokenize_util.cpp otherarch/sdcpp/version.cpp otherarch/sdcpp/thirdparty/zip.c build-info.h ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o console.o ggml-backend_default.o ggml-backend-reg_default.o ggml-repack.o $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 whispermain: otherarch/whispercpp/main.cpp otherarch/whispercpp/whisper.cpp build-info.h ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o console.o ggml-backend_default.o ggml-backend-reg_default.o ggml-repack.o $(OBJS_FULL) $(OBJS)

gpttype_adapter.cpp

@@ -152,7 +152,7 @@ static int delayed_generated_tokens_limit = 0;
 std::deque<std::string> delayed_generated_tokens; //for use with antislop sampling
 static std::map<int,std::vector<int>> antislop_banned_token_ids; //first is the npast position, second is the array of banned ids at that index
 
-const int savestate_limit = 4;
+const int savestate_limit = 5;
 static savestate_data savestates[savestate_limit];
 
 inline int kcpp_cpu_has_blas(void) {

koboldcpp.py

@@ -53,7 +53,7 @@ default_draft_amount = 8
 default_ttsmaxlen = 4096
 default_visionmaxres = 1024
 net_save_slots = 12
-savestate_limit = 4 #savestate slots
+savestate_limit = 5 #savestate slots
 default_vae_tile_threshold = 768
 default_native_ctx = 16384
 overridekv_max = 4

tools/fit-params/fit-params.cpp

@@ -9,6 +9,8 @@
 #if defined(_MSC_VER)
 #pragma warning(disable: 4244 4267) // possible loss of data
 #endif
+#include <chrono>
+#include <thread>
 
 int main(int argc, char ** argv) {
     common_params params;
@@ -27,6 +29,7 @@ int main(int argc, char ** argv) {
         params.verbosity >= 4 ? GGML_LOG_LEVEL_DEBUG : GGML_LOG_LEVEL_ERROR);
 
     LOG_INF("Printing fitted CLI arguments to stdout...\n");
+    std::this_thread::sleep_for(std::chrono::seconds(1));
     std::cout << "-c "    << cparams.n_ctx;
     std::cout << " -ngl " << mparams.n_gpu_layers;