fixed incorrect makefile flags

Makefile

@@ -430,6 +430,7 @@ ggml_v4_vulkan.o: ggml/src/ggml.c ggml/include/ggml.h
 ggml_v4_vulkan_noavx2.o: ggml/src/ggml.c ggml/include/ggml.h
 	$(CC)  $(FASTCFLAGS) $(SIMPLECFLAGS) $(VULKAN_FLAGS) -c $< -o $@
 
+# cpu and clblast separated
 ggml-cpu.o: ggml/src/ggml-cpu.c ggml/include/ggml-cpu.h
 	$(CC)  $(FASTCFLAGS) $(FULLCFLAGS) -c $< -o $@
 ggml-cpu_v4_failsafe.o: ggml/src/ggml-cpu.c ggml/include/ggml-cpu.h
@@ -563,7 +564,7 @@ whispercpp_cublas.o: otherarch/whispercpp/whisper_adapter.cpp
 	$(CXX) $(CXXFLAGS) $(CUBLAS_FLAGS) $(HIPFLAGS) -c $< -o $@
 
 # idiotic "for easier compilation"
-GPTTYPE_ADAPTER = gpttype_adapter.cpp otherarch/llama_v2.cpp otherarch/llama_v3.cpp src/llama.cpp src/llama-grammar.cpp src/llama-sampling.cpp src/llama-vocab.cpp otherarch/utils.cpp otherarch/gptj_v1.cpp otherarch/gptj_v2.cpp otherarch/gptj_v3.cpp otherarch/gpt2_v1.cpp otherarch/gpt2_v2.cpp otherarch/gpt2_v3.cpp otherarch/rwkv_v2.cpp otherarch/rwkv_v3.cpp otherarch/neox_v2.cpp otherarch/neox_v3.cpp otherarch/mpt_v3.cpp ggml/include/ggml.h ggml/include/ggml-cuda.h include/llama.h otherarch/llama-util.h
+GPTTYPE_ADAPTER = gpttype_adapter.cpp otherarch/llama_v2.cpp otherarch/llama_v3.cpp src/llama.cpp src/llama-grammar.cpp src/llama-sampling.cpp src/llama-vocab.cpp otherarch/utils.cpp otherarch/gptj_v1.cpp otherarch/gptj_v2.cpp otherarch/gptj_v3.cpp otherarch/gpt2_v1.cpp otherarch/gpt2_v2.cpp otherarch/gpt2_v3.cpp otherarch/rwkv_v2.cpp otherarch/rwkv_v3.cpp otherarch/neox_v2.cpp otherarch/neox_v3.cpp otherarch/mpt_v3.cpp ggml/include/ggml.h ggml/include/ggml-cpu.h ggml/include/ggml-cuda.h include/llama.h otherarch/llama-util.h
 gpttype_adapter_failsafe.o: $(GPTTYPE_ADAPTER)
 	$(CXX) $(CXXFLAGS) $(FAILSAFE_FLAGS) -c $< -o $@
 gpttype_adapter.o: $(GPTTYPE_ADAPTER)
@@ -643,7 +644,7 @@ koboldcpp_clblast_noavx2:
 endif
 
 ifdef CUBLAS_BUILD
-koboldcpp_cublas: ggml_v4_cublas.o ggml.o ggml_v3_cublas.o ggml_v2_cublas.o ggml_v1.o expose.o gpttype_adapter_cublas.o sdcpp_cublas.o whispercpp_cublas.o llavaclip_cublas.o llava.o ggml-backend_cublas.o $(CUBLAS_OBJS) $(OBJS_FULL) $(OBJS)
+koboldcpp_cublas: ggml_v4_cublas.o ggml-cpu.o ggml_v3_cublas.o ggml_v2_cublas.o ggml_v1.o expose.o gpttype_adapter_cublas.o sdcpp_cublas.o whispercpp_cublas.o llavaclip_cublas.o llava.o ggml-backend_cublas.o $(CUBLAS_OBJS) $(OBJS_FULL) $(OBJS)
 	$(CUBLAS_BUILD)
 else
 koboldcpp_cublas:
@@ -651,7 +652,7 @@ koboldcpp_cublas:
 endif
 
 ifdef HIPBLAS_BUILD
-koboldcpp_hipblas: ggml_v4_cublas.o ggml.o ggml_v3_cublas.o ggml_v2_cublas.o ggml_v1.o expose.o gpttype_adapter_cublas.o sdcpp_cublas.o whispercpp_cublas.o llavaclip_cublas.o llava.o ggml-backend_cublas.o $(HIP_OBJS) $(OBJS_FULL) $(OBJS)
+koboldcpp_hipblas: ggml_v4_cublas.o ggml-cpu.o ggml_v3_cublas.o ggml_v2_cublas.o ggml_v1.o expose.o gpttype_adapter_cublas.o sdcpp_cublas.o whispercpp_cublas.o llavaclip_cublas.o llava.o ggml-backend_cublas.o $(HIP_OBJS) $(OBJS_FULL) $(OBJS)
 	$(HIPBLAS_BUILD)
 else
 koboldcpp_hipblas:
@@ -662,7 +663,7 @@ ifdef VULKAN_BUILD
 koboldcpp_vulkan: ggml_v4_vulkan.o ggml-cpu.o ggml_v3.o ggml_v2.o ggml_v1.o expose.o gpttype_adapter_vulkan.o ggml-vulkan.o sdcpp_vulkan.o whispercpp_default.o llavaclip_vulkan.o llava.o ggml-backend_vulkan.o $(OBJS_FULL) $(OBJS)
 	$(VULKAN_BUILD)
 ifdef NOAVX2_BUILD
-koboldcpp_vulkan_noavx2: ggml_v4_vulkan_noavx2.o ggml_v4_noavx2.o ggml_v3_noavx2.o ggml_v2_noavx2.o ggml_v1_failsafe.o expose.o gpttype_adapter_vulkan_noavx2.o ggml-vulkan.o sdcpp_vulkan.o whispercpp_default.o llavaclip_vulkan.o llava.o ggml-backend_vulkan.o $(OBJS_SIMPLE) $(OBJS)
+koboldcpp_vulkan_noavx2: ggml_v4_vulkan_noavx2.o ggml-cpu_v4_noavx2.o ggml_v3_noavx2.o ggml_v2_noavx2.o ggml_v1_failsafe.o expose.o gpttype_adapter_vulkan_noavx2.o ggml-vulkan.o sdcpp_vulkan.o whispercpp_default.o llavaclip_vulkan.o llava.o ggml-backend_vulkan.o $(OBJS_SIMPLE) $(OBJS)
 	$(VULKAN_BUILD)
 else
 koboldcpp_vulkan_noavx2:

ggml/src/ggml-vulkan.cpp

@@ -1035,7 +1035,7 @@ static uint32_t find_properties(const vk::PhysicalDeviceMemoryProperties* mem_pr
 static vk_buffer ggml_vk_create_buffer(vk_device& device, size_t size, vk::MemoryPropertyFlags req_flags, vk::MemoryPropertyFlags fallback_flags = vk::MemoryPropertyFlags(0)) {
     VK_LOG_DEBUG("ggml_vk_create_buffer(" << device->name << ", " << size << ", " << to_string(req_flags) << ", " << to_string(fallback_flags) << ")");
     if (size > device->max_memory_allocation_size) {
-        printf("\nWARNING: Requested buffer size exceeds device memory allocation limit!\n");
+        printf("\nWARNING: Requested buffer size (%zu) exceeds device memory allocation limit (%zu)!\n",size,device->max_memory_allocation_size);
     }
 
     std::lock_guard<std::mutex> guard(device->mutex);

koboldcpp.py

@@ -45,7 +45,7 @@ maxhordelen = 400
 modelbusy = threading.Lock()
 requestsinqueue = 0
 defaultport = 5001
-KcppVersion = "1.77"
+KcppVersion = "1.78"
 showdebug = True
 guimode = False
 showsamplerwarning = True
@@ -2867,7 +2867,7 @@ def show_gui():
     makeslider(quick_tab, "Context Size:", contextsize_text, context_var, 0, len(contextsize_text)-1, 30, width=280, set=5,tooltip="What is the maximum context size to support. Model specific. You cannot exceed it.\nLarger contexts require more memory, and not all models support it.")
 
     # load model
-    makefileentry(quick_tab, "Model:", "Select GGUF or GGML Model File", model_var, 40, 280, onchoosefile=on_picked_model_file,tooltiptxt="Select a GGUF or GGML model file on disk to be loaded.")
+    makefileentry(quick_tab, "GGUF Model:", "Select GGUF or GGML Model File", model_var, 40, 280, onchoosefile=on_picked_model_file,tooltiptxt="Select a GGUF or GGML model file on disk to be loaded.")
     model_var.trace("w", gui_changed_modelfile)
 
     # Hardware Tab

otherarch/sdcpp/model.cpp

@@ -14,6 +14,7 @@
 #include "ggml-alloc.h"
 #include "ggml-backend.h"
 #include "ggml.h"
+#include "ggml-cpu.h"
 
 #include "stable-diffusion.h"