Merge branch 'concedo_experimental' into frankenstein_11

Author: Nexesenex

.github/labeler.yml

@@ -0,0 +1,78 @@
+# https://github.com/actions/labeler
+
+SYCL:
+    - changed-files:
+        - any-glob-to-any-file:
+            - ggml-sycl.h
+            - ggml-sycl.cpp
+            - README-sycl.md
+Nvidia GPU:
+    - changed-files:
+        - any-glob-to-any-file:
+            - ggml-cuda/**
+Vulkan:
+    - changed-files:
+        - any-glob-to-any-file:
+            - ggml_vk_generate_shaders.py
+            - ggml-vulkan*
+documentation:
+    - changed-files:
+        - any-glob-to-any-file:
+            - docs/**
+            - media/**
+testing:
+    - changed-files:
+        - any-glob-to-any-file:
+            - tests/**
+build:
+    - changed-files:
+        - any-glob-to-any-file:
+            - cmake/**
+            - CMakeLists.txt
+            - CMakePresets.json
+            - codecov.yml
+examples:
+    - changed-files:
+        - any-glob-to-any-file: examples/**
+devops:
+    - changed-files:
+        - any-glob-to-any-file:
+            - .devops/**
+            - .github/**
+            - ci/**
+python:
+    - changed-files:
+        - any-glob-to-any-file:
+            - "**/*.py"
+            - requirements/**
+            - gguf-py/**
+            - .flake8
+script:
+    - changed-files:
+        - any-glob-to-any-file:
+            - scripts/**
+android:
+    - changed-files:
+        - any-glob-to-any-file:
+            - examples/llama.android/**
+server:
+    - changed-files:
+        - any-glob-to-any-file:
+            - examples/server/**
+ggml:
+    - changed-files:
+        - any-glob-to-any-file:
+            - ggml.c
+            - ggml.h
+            - ggml-*.c
+            - ggml-*.h
+            - ggml-cuda/**
+nix:
+    - changed-files:
+        - any-glob-to-any-file:
+            - "**/*.nix"
+            - .github/workflows/nix-*.yml
+            - .devops/nix/nixpkgs-instances.nix
+embedding:
+    - changed-files:
+        - any-glob-to-any-file: examples/embedding/

.github/workflows/labeler.yml

@@ -0,0 +1,17 @@
+name: "Pull Request Labeler"
+on:
+- pull_request_target
+
+jobs:
+  labeler:
+    permissions:
+      contents: read
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        repository: "ggerganov/llama.cpp"
+    - uses: actions/labeler@v5
+      with:
+        configuration-path: '.github/labeler.yml'

CMakeLists.txt

@@ -428,6 +428,8 @@ set_target_properties(ggml_v3 PROPERTIES POSITION_INDEPENDENT_CODE ON)
 add_library(common2
             common/common.cpp
             common/common.h
+            common/sampling.cpp
+            common/sampling.h
             common/grammar-parser.h
             common/grammar-parser.cpp
             examples/llava/llava.cpp

Makefile

@@ -62,9 +62,9 @@ endif
 CUBLASLD_FLAGS =
 CUBLAS_OBJS =
 
-OBJS_FULL += ggml-alloc.o ggml-quants.o unicode.o unicode-data.o sgemm.o common.o grammar-parser.o
-OBJS_SIMPLE += ggml-alloc.o ggml-quants_noavx2.o unicode.o unicode-data.o sgemm_noavx2.o common.o grammar-parser.o
-OBJS_FAILSAFE += ggml-alloc.o ggml-quants_failsafe.o unicode.o unicode-data.o sgemm_failsafe.o common.o grammar-parser.o
+OBJS_FULL += ggml-alloc.o ggml-quants.o unicode.o unicode-data.o sgemm.o common.o sampling.o grammar-parser.o
+OBJS_SIMPLE += ggml-alloc.o ggml-quants_noavx2.o unicode.o unicode-data.o sgemm_noavx2.o common.o sampling.o grammar-parser.o
+OBJS_FAILSAFE += ggml-alloc.o ggml-quants_failsafe.o unicode.o unicode-data.o sgemm_failsafe.o common.o sampling.o grammar-parser.o
 
 #lets try enabling everything
 CFLAGS   += -pthread -s -Wno-deprecated -Wno-deprecated-declarations
@@ -497,6 +497,8 @@ llama.o: llama.cpp ggml.h ggml-alloc.h ggml-backend.h ggml-cuda.h ggml-metal.h l
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 common.o: common/common.cpp common/common.h common/log.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
+sampling.o: common/sampling.cpp common/common.h common/sampling.h common/log.h
+	$(CXX) $(CXXFLAGS) -c $< -o $@
 console.o: common/console.cpp common/console.h
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 grammar-parser.o: common/grammar-parser.cpp common/grammar-parser.h

README.md

@@ -58,7 +58,7 @@ when you can't use the precompiled binary directly, we provide an automated buil
   - For Arch Linux: Install `cblas` `openblas` and `clblast`.
   - For Debian: Install `libclblast-dev` and `libopenblas-dev`.
 - You can attempt a CuBLAS build with `LLAMA_CUBLAS=1`. You will need CUDA Toolkit installed. Some have also reported success with the CMake file, though that is more for windows.
-- For a full featured build (all backends), do `make LLAMA_OPENBLAS=1 LLAMA_CLBLAST=1 LLAMA_CUBLAS=1 LLAMA_VULKAN=1`
+- For a full featured build (all backends), do `make LLAMA_OPENBLAS=1 LLAMA_CLBLAST=1 LLAMA_CUBLAS=1 LLAMA_VULKAN=1`. (Note that `LLAMA_CUBLAS=1` will not work on windows, you need visual studio)
 - After all binaries are built, you can run the python script with the command `koboldcpp.py [ggml_model.bin] [port]`
 
 - Note: Many OSX users have found that the using Accelerate is actually faster than OpenBLAS. To try, you may wish to run with `--noblas` and compare speeds.