Merge branch 'remoteManagement' into crokeso

Author: Nexesenex

.editorconfig

@@ -21,23 +21,23 @@ indent_style = tab
 [prompts/*.txt]
 insert_final_newline = unset
 
-[examples/server/public/*]
+[tools/server/public/*]
 indent_size = 2
 
-[examples/server/public/deps_*]
+[tools/server/public/deps_*]
 trim_trailing_whitespace = unset
 indent_style = unset
 indent_size = unset
 
-[examples/server/deps_*]
+[tools/server/deps_*]
 trim_trailing_whitespace = unset
 indent_style = unset
 indent_size = unset
 
 [examples/llama.swiftui/llama.swiftui.xcodeproj/*]
 indent_style = tab
 
-[examples/cvector-generator/*.txt]
+[tools/cvector-generator/*.txt]
 trim_trailing_whitespace = unset
 insert_final_newline = unset
 

.gitignore

@@ -92,8 +92,6 @@ ppl-*.txt
 qnt-*.txt
 perf-*.txt
 
-examples/jeopardy/results.txt
-
 poetry.lock
 poetry.toml
 

CMakeLists.txt

@@ -847,59 +847,59 @@ add_library(common2
             common/common.h
             common/sampling.cpp
             common/sampling.h
-            examples/llava/llava.cpp
-            examples/llava/llava.h
-            examples/llava/clip.cpp
-            examples/llava/clip.h
+            tools/llava/llava.cpp
+            tools/llava/llava.h
+            tools/llava/clip.cpp
+            tools/llava/clip.h
             src/unicode.h
             src/unicode.cpp
             src/unicode-data.cpp
             otherarch/utils.cpp
             otherarch/utils.h)
-target_include_directories(common2 PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./examples ./common)
+target_include_directories(common2 PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./tools ./common)
 target_compile_features(common2 PUBLIC cxx_std_17) # don't bump
 target_link_libraries(common2 PRIVATE ggml ${LLAMA_EXTRA_LIBS})
 set_target_properties(common2 PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 add_library(sdtype_adapter
             otherarch/sdcpp/sdtype_adapter.cpp)
-target_include_directories(sdtype_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./examples ./common)
+target_include_directories(sdtype_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./tools ./common)
 target_compile_features(sdtype_adapter PUBLIC cxx_std_17) # don't bump
 target_link_libraries(sdtype_adapter PRIVATE common2 ggml ${LLAMA_EXTRA_LIBS})
 set_target_properties(sdtype_adapter PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 add_library(whisper_adapter
             otherarch/whispercpp/whisper_adapter.cpp)
-target_include_directories(whisper_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/whispercpp ./examples ./common)
+target_include_directories(whisper_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/whispercpp ./tools ./common)
 target_compile_features(whisper_adapter PUBLIC cxx_std_17) # don't bump
 target_link_libraries(whisper_adapter PRIVATE common2 ggml ${LLAMA_EXTRA_LIBS})
 set_target_properties(whisper_adapter PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 add_library(tts_adapter
             otherarch/tts_adapter.cpp)
-target_include_directories(tts_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./examples ./common)
+target_include_directories(tts_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./tools ./common)
 target_compile_features(tts_adapter PUBLIC cxx_std_17) # don't bump
 target_link_libraries(tts_adapter PRIVATE common2 ggml ${LLAMA_EXTRA_LIBS})
 set_target_properties(tts_adapter PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 add_library(embeddings_adapter
             otherarch/embeddings_adapter.cpp)
-target_include_directories(embeddings_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./examples ./common)
+target_include_directories(embeddings_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./tools ./common)
 target_compile_features(embeddings_adapter PUBLIC cxx_std_17) # don't bump
 target_link_libraries(embeddings_adapter PRIVATE common2 ggml ${LLAMA_EXTRA_LIBS})
 set_target_properties(embeddings_adapter PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 add_library(gpttype_adapter
             gpttype_adapter.cpp)
-target_include_directories(gpttype_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./examples ./common)
+target_include_directories(gpttype_adapter PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./tools ./common)
 target_compile_features(gpttype_adapter PUBLIC cxx_std_17) # don't bump
 target_link_libraries(gpttype_adapter PRIVATE common2 ggml ggml_v1 ggml_v2 ggml_v3 ${LLAMA_EXTRA_LIBS})
 set_target_properties(gpttype_adapter PROPERTIES POSITION_INDEPENDENT_CODE ON)
 
 if (LLAMA_CUBLAS)
     set(TARGET koboldcpp_cublas)
     add_library(${TARGET} SHARED expose.cpp expose.h)
-    target_include_directories(${TARGET} PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./examples ./common)
+    target_include_directories(${TARGET} PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./tools ./common)
     target_compile_features(${TARGET} PUBLIC cxx_std_17) # don't bump
     set_target_properties(${TARGET} PROPERTIES PREFIX "")
     set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME "koboldcpp_cublas")
@@ -919,7 +919,7 @@ endif()
 if (LLAMA_HIPBLAS)
     set(TARGET koboldcpp_hipblas)
     add_library(${TARGET} SHARED expose.cpp expose.h)
-    target_include_directories(${TARGET} PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./examples ./common)
+    target_include_directories(${TARGET} PUBLIC . ./ggml/include ./ggml/src ./ggml/src/ggml-cpu ./include ./otherarch ./otherarch/tools ./otherarch/sdcpp ./otherarch/sdcpp/thirdparty ./tools ./common)
     target_compile_features(${TARGET} PUBLIC cxx_std_17) # don't bump
     set_target_properties(${TARGET} PROPERTIES PREFIX "")
     set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME "koboldcpp_hipblas")

Makefile

@@ -584,7 +584,7 @@ sgemm_failsafe.o: ggml/src/ggml-cpu/llamafile/sgemm.cpp ggml/src/ggml-cpu/llamaf
 #there's no intrinsics or special gpu ops used here, so we can have a universal object
 ggml-alloc.o: ggml/src/ggml-alloc.c ggml/include/ggml.h ggml/include/ggml-alloc.h
 	$(CC)  $(CFLAGS) -c $< -o $@
-llava.o: examples/llava/llava.cpp examples/llava/llava.h
+llava.o: tools/llava/llava.cpp tools/llava/llava.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 unicode.o: src/unicode.cpp src/unicode.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
@@ -614,11 +614,11 @@ ggml-backend-reg_vulkan.o: ggml/src/ggml-backend-reg.cpp ggml/src/ggml-backend-i
 	$(CXX)  $(CXXFLAGS) $(VULKAN_FLAGS) -c $< -o $@
 ggml-backend-reg_cublas.o: ggml/src/ggml-backend-reg.cpp ggml/src/ggml-backend-impl.h ggml/include/ggml.h ggml/include/ggml-backend.h ggml/include/ggml-cpu.h
 	$(CXX)  $(CXXFLAGS) $(CUBLAS_FLAGS) $(HIPFLAGS) -c $< -o $@
-llavaclip_default.o: examples/llava/clip.cpp examples/llava/clip.h
+llavaclip_default.o: tools/llava/clip.cpp tools/llava/clip.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
-llavaclip_cublas.o: examples/llava/clip.cpp examples/llava/clip.h
+llavaclip_cublas.o: tools/llava/clip.cpp tools/llava/clip.h
 	$(CXX) $(CXXFLAGS) $(CUBLAS_FLAGS) $(HIPFLAGS) -c $< -o $@
-llavaclip_vulkan.o: examples/llava/clip.cpp examples/llava/clip.h
+llavaclip_vulkan.o: tools/llava/clip.cpp tools/llava/clip.h
 	$(CXX) $(CXXFLAGS) $(VULKAN_FLAGS) -c $< -o $@
 
 #this is only used for accelerate
@@ -737,17 +737,17 @@ cleanocuda:
 	# rm -vrf ggml/src/ggml-cuda/template-instances/*.o
 
 # useful tools
-main: examples/main/main.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
+main: tools/main/main.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -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/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 $(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 $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
-ttsmain: examples/tts/tts.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
+ttsmain: tools/tts/tts.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
-gguf-split: examples/gguf-split/gguf-split.cpp ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o build-info.h llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
+gguf-split: tools/gguf-split/gguf-split.cpp ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o build-info.h llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
-mtmd-cli: examples/llava/mtmd-cli.cpp examples/llava/mtmd.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
+mtmd-cli: tools/llava/mtmd-cli.cpp tools/llava/mtmd.cpp common/arg.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 llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL) $(OBJS)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h,$^) -o $@ $(LDFLAGS)
 
 ggml/src/ggml-vulkan-shaders.cpp:
@@ -887,7 +887,7 @@ koboldcpp_vulkan_noavx2:
 endif
 
 # tools
-quantize_gguf: examples/quantize/quantize.cpp ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
+quantize_gguf: tools/quantize/quantize.cpp ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
 quantize_gptj: otherarch/tools/gptj_quantize.cpp otherarch/tools/common-ggml.cpp ggml_v3.o ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
@@ -897,7 +897,7 @@ quantize_neox: otherarch/tools/neox_quantize.cpp otherarch/tools/common-ggml.cpp
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
 quantize_mpt: otherarch/tools/mpt_quantize.cpp otherarch/tools/common-ggml.cpp ggml_v3.o ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o llavaclip_default.o llava.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
-quantize_clip: examples/llava/clip.cpp examples/llava/clip.h examples/llava/quantclip.cpp ggml_v3.o ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
+quantize_clip: tools/llava/clip.cpp tools/llava/clip.h tools/quantclip.cpp ggml_v3.o ggml.o ggml-cpu.o ggml-ops.o ggml-vec.o ggml-binops.o ggml-unops.o llama.o ggml-backend_default.o ggml-backend-reg_default.o $(OBJS_FULL)
 	$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
 
 #window simple clinfo

common/arg.cpp

@@ -2213,14 +2213,14 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
     ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_NO_CONT_BATCHING"));
     add_opt(common_arg(
         {"--mmproj"}, "FILE",
-        "path to a multimodal projector file. see examples/llava/README.md",
+        "path to a multimodal projector file. see tools/llava/README.md",
         [](common_params & params, const std::string & value) {
             params.mmproj.path = value;
         }
     ).set_examples(mmproj_examples));
     add_opt(common_arg(
         {"--mmproj-url"}, "URL",
-        "URL to a multimodal projector file. see examples/llava/README.md",
+        "URL to a multimodal projector file. see tools/llava/README.md",
         [](common_params & params, const std::string & value) {
             params.mmproj.url = value;
         }

common/common.h

@@ -336,7 +336,7 @@ struct common_params {
 
     common_conversation_mode conversation_mode = COMMON_CONVERSATION_MODE_AUTO;
 
-    // multimodal models (see examples/llava)
+    // multimodal models (see tools/llava)
     struct common_params_model mmproj;
     bool mmproj_use_gpu = true;     // use GPU for multimodal model
     bool no_mmproj = false;         // explicitly disable multimodal model
@@ -410,8 +410,8 @@ struct common_params {
     int n_pca_batch = 100;
     int n_pca_iterations = 1000;
     dimre_method cvector_dimre_method = DIMRE_METHOD_PCA;
-    std::string cvector_positive_file = "examples/cvector-generator/positive.txt";
-    std::string cvector_negative_file = "examples/cvector-generator/negative.txt";
+    std::string cvector_positive_file = "tools/cvector-generator/positive.txt";
+    std::string cvector_negative_file = "tools/cvector-generator/negative.txt";
 
     bool spm_infill = false; // suffix/prefix/middle pattern for infill