Merge pull request #204 from ggerganov/master

b3284

.devops/nix/package.nix

@@ -17,19 +17,18 @@
   rocmPackages,
   vulkan-headers,
   vulkan-loader,
-  clblast,
+  curl,
   useBlas ? builtins.all (x: !x) [
     useCuda
     useMetalKit
-    useOpenCL
     useRocm
     useVulkan
   ] && blas.meta.available,
   useCuda ? config.cudaSupport,
-  useMetalKit ? stdenv.isAarch64 && stdenv.isDarwin && !useOpenCL,
+  useMetalKit ? stdenv.isAarch64 && stdenv.isDarwin,
   useMpi ? false, # Increases the runtime closure size by ~700M
-  useOpenCL ? false,
   useRocm ? config.rocmSupport,
+  enableCurl ? true,
   useVulkan ? false,
   llamaVersion ? "0.0.0", # Arbitrary version, substituted by the flake
 
@@ -56,7 +55,6 @@ let
     ++ lib.optionals useCuda [ "CUDA" ]
     ++ lib.optionals useMetalKit [ "MetalKit" ]
     ++ lib.optionals useMpi [ "MPI" ]
-    ++ lib.optionals useOpenCL [ "OpenCL" ]
     ++ lib.optionals useRocm [ "ROCm" ]
     ++ lib.optionals useVulkan [ "Vulkan" ];
 
@@ -198,19 +196,19 @@ effectiveStdenv.mkDerivation (
       optionals effectiveStdenv.isDarwin darwinBuildInputs
       ++ optionals useCuda cudaBuildInputs
       ++ optionals useMpi [ mpi ]
-      ++ optionals useOpenCL [ clblast ]
       ++ optionals useRocm rocmBuildInputs
       ++ optionals useBlas [ blas ]
-      ++ optionals useVulkan vulkanBuildInputs;
+      ++ optionals useVulkan vulkanBuildInputs
+      ++ optionals enableCurl [ curl ];
 
     cmakeFlags =
       [
         (cmakeBool "LLAMA_BUILD_SERVER" true)
         (cmakeBool "BUILD_SHARED_LIBS" (!enableStatic))
         (cmakeBool "CMAKE_SKIP_BUILD_RPATH" true)
+        (cmakeBool "LLAMA_CURL" enableCurl)
         (cmakeBool "GGML_NATIVE" false)
         (cmakeBool "GGML_BLAS" useBlas)
-        (cmakeBool "GGML_CLBLAST" useOpenCL)
         (cmakeBool "GGML_CUDA" useCuda)
         (cmakeBool "GGML_HIPBLAS" useRocm)
         (cmakeBool "GGML_METAL" useMetalKit)
@@ -254,7 +252,6 @@ effectiveStdenv.mkDerivation (
         useCuda
         useMetalKit
         useMpi
-        useOpenCL
         useRocm
         useVulkan
         ;
@@ -281,7 +278,7 @@ effectiveStdenv.mkDerivation (
       # Configurations we don't want even the CI to evaluate. Results in the
       # "unsupported platform" messages. This is mostly a no-op, because
       # cudaPackages would've refused to evaluate anyway.
-      badPlatforms = optionals (useCuda || useOpenCL) lib.platforms.darwin;
+      badPlatforms = optionals useCuda lib.platforms.darwin;
 
       # Configurations that are known to result in build failures. Can be
       # overridden by importing Nixpkgs with `allowBroken = true`.

.github/ISSUE_TEMPLATE/config.yml

@@ -9,5 +9,3 @@ contact_links:
   - name: Want to contribute?
     url: https://github.com/ggerganov/llama.cpp/wiki/contribute
     about: Head to the contribution guide page of the wiki for areas you can help with
-
-

.github/workflows/build.yml

@@ -799,6 +799,7 @@ jobs:
           7z x "-o${env:RUNNER_TEMP}" $env:RUNNER_TEMP/sde.tar
           $sde = $(join-path $env:RUNNER_TEMP sde-external-${env:SDE_VERSION}-win/sde.exe)
           cd build
+          $env:LLAMA_SKIP_TESTS_SLOW_ON_EMULATOR = 1
           & $sde -future -- ctest -L main -C Release --verbose --timeout 900
 
       - name: Determine tag name

CMakeLists.txt

@@ -79,8 +79,15 @@ set(GGML_SANITIZE_ADDRESS   ${LLAMA_SANITIZE_ADDRESS})
 set(GGML_SANITIZE_UNDEFINED ${LLAMA_SANITIZE_UNDEFINED})
 set(GGML_ALL_WARNINGS       ${LLAMA_ALL_WARNINGS})
 set(GGML_FATAL_WARNINGS     ${LLAMA_FATAL_WARNINGS})
-set(GGML_LLAMAFILE          ON)
-set(GGML_CUDA_USE_GRAPHS    ON)
+
+# change the default for these ggml options
+if (NOT DEFINED GGML_LLAMAFILE)
+    set(GGML_LLAMAFILE ON)
+endif()
+
+if (NOT DEFINED GGML_CUDA_USE_GRAPHS)
+    set(GGML_CUDA_USE_GRAPHS ON)
+endif()
 
 # transition helpers
 function (llama_option_depr TYPE OLD NEW)

CMakePresets.json

@@ -19,6 +19,7 @@
         "cacheVariables": {
             "CMAKE_EXPORT_COMPILE_COMMANDS": "ON",
             "CMAKE_CXX_COMPILER": "icx",
+            "CMAKE_C_COMPILER": "cl",
             "GGML_SYCL": "ON",
             "CMAKE_INSTALL_RPATH": "$ORIGIN;$ORIGIN/.."
         }

Makefile

@@ -62,6 +62,11 @@ TEST_TARGETS = \
 	tests/test-tokenizer-1-bpe \
 	tests/test-tokenizer-1-spm
 
+# Legacy build targets that were renamed in #7809, but should still be removed when the project is cleaned
+LEGACY_TARGETS = main quantize quantize-stats perplexity imatrix embedding vdot q8dot train-text-from-scratch convert-llama2c-to-ggml \
+	simple batched batched-bench save-load-state server gguf gguf-split eval-callback llama-bench libllava.a llava-cli baby-llama \
+	retrieval speculative infill tokenize benchmark-matmult parallel finetune export-lora lookahead lookup passkey gritlm
+
 # Deprecation aliases
 ifdef LLAMA_CUBLAS
 $(error LLAMA_CUBLAS is removed. Use GGML_CUDA instead.)
@@ -1086,6 +1091,7 @@ clean:
 	rm -vrf ggml/src/ggml-cuda/template-instances/*.o
 	rm -rvf $(BUILD_TARGETS)
 	rm -rvf $(TEST_TARGETS)
+	rm -rvf $(LEGACY_TARGETS)
 	find examples pocs -type f -name "*.o" -delete
 
 #

README.md

@@ -108,6 +108,7 @@ Typically finetunes of the base models below are supported as well.
 - [X] [Falcon](https://huggingface.co/models?search=tiiuae/falcon)
 - [X] [Chinese LLaMA / Alpaca](https://github.com/ymcui/Chinese-LLaMA-Alpaca) and [Chinese LLaMA-2 / Alpaca-2](https://github.com/ymcui/Chinese-LLaMA-Alpaca-2)
 - [X] [Vigogne (French)](https://github.com/bofenghuang/vigogne)
+- [X] [BERT](https://github.com/ggerganov/llama.cpp/pull/5423)
 - [X] [Koala](https://bair.berkeley.edu/blog/2023/04/03/koala/)
 - [X] [Baichuan 1 & 2](https://huggingface.co/models?search=baichuan-inc/Baichuan) + [derivations](https://huggingface.co/hiyouga/baichuan-7b-sft)
 - [X] [Aquila 1 & 2](https://huggingface.co/models?search=BAAI/Aquila)
@@ -217,6 +218,11 @@ Unless otherwise noted these projects are open-source with permissive licensing:
 **Tools:**
 
 - [akx/ggify](https://github.com/akx/ggify) – download PyTorch models from HuggingFace Hub and convert them to GGML
+- [crashr/gppm](https://github.com/crashr/gppm) – launch llama.cpp instances utilizing NVIDIA Tesla P40 or P100 GPUs with reduced idle power consumption
+
+**Infrastructure:**
+
+- [Paddler](https://github.com/distantmagic/paddler) - Stateful load balancer custom-tailored for llama.cpp
 
 ---
 

common/common.cpp

@@ -1014,16 +1014,23 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
     }
     if (arg == "--in-prefix-bos") {
         params.input_prefix_bos = true;
+        params.enable_chat_template = false;
         return true;
     }
     if (arg == "--in-prefix") {
         CHECK_ARG
         params.input_prefix = argv[i];
+        params.enable_chat_template = false;
         return true;
     }
     if (arg == "--in-suffix") {
         CHECK_ARG
         params.input_suffix = argv[i];
+        params.enable_chat_template = false;
+        return true;
+    }
+    if (arg == "--spm-infill") {
+        params.spm_infill = true;
         return true;
     }
     if (arg == "--grammar") {
@@ -1402,13 +1409,15 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
                                                                         "halt generation at PROMPT, return control in interactive mode\n"
                                                                         "can be specified more than once for multiple prompts" });
     options.push_back({ "main",        "-sp,   --special",              "special tokens output enabled (default: %s)", params.special ? "true" : "false" });
-    options.push_back({ "main",        "-cnv,  --conversation",         "run in conversation mode (does not print special tokens and suffix/prefix) (default: %s)", params.conversation ? "true" : "false" });
+    options.push_back({ "main",        "-cnv,  --conversation",         "run in conversation mode (does not print special tokens and suffix/prefix, use default chat template) (default: %s)", params.conversation ? "true" : "false" });
     options.push_back({ "main infill", "-i,    --interactive",          "run in interactive mode (default: %s)", params.interactive ? "true" : "false" });
     options.push_back({ "main infill", "-if,   --interactive-first",    "run in interactive mode and wait for input right away (default: %s)", params.interactive_first ? "true" : "false" });
     options.push_back({ "main infill", "-mli,  --multiline-input",      "allows you to write or paste multiple lines without ending each in '\\'" });
     options.push_back({ "main infill", "       --in-prefix-bos",        "prefix BOS to user inputs, preceding the `--in-prefix` string" });
     options.push_back({ "main infill", "       --in-prefix STRING",     "string to prefix user inputs with (default: empty)" });
     options.push_back({ "main infill", "       --in-suffix STRING",     "string to suffix after user inputs with (default: empty)" });
+    options.push_back({ "server infill",
+                                       "       --spm-infill",           "use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: %s)", params.spm_infill ? "enabled" : "disabled" });
 
     options.push_back({ "sampling" });
     options.push_back({ "*",           "       --samplers SAMPLERS",    "samplers that will be used for generation in the order, separated by \';\'\n"
@@ -2662,12 +2671,19 @@ std::string llama_chat_format_single(const struct llama_model * model,
         const std::vector<llama_chat_msg> & past_msg,
         const llama_chat_msg & new_msg,
         bool add_ass) {
+    std::ostringstream ss;
     auto fmt_past_msg = llama_chat_apply_template(model, tmpl, past_msg, false);
     std::vector<llama_chat_msg> chat_new(past_msg);
+    // if the past_msg ends with a newline, we must preserve it in the formatted version
+    if (add_ass && !fmt_past_msg.empty() && fmt_past_msg.back() == '\n') {
+        ss << "\n";
+    };
+    // format chat with new_msg
     chat_new.push_back(new_msg);
     auto fmt_new_msg = llama_chat_apply_template(model, tmpl, chat_new, add_ass);
-    auto formatted = fmt_new_msg.substr(fmt_past_msg.size(), fmt_new_msg.size() - fmt_past_msg.size());
-    return formatted;
+    // get the diff part
+    ss << fmt_new_msg.substr(fmt_past_msg.size(), fmt_new_msg.size() - fmt_past_msg.size());
+    return ss.str();
 }
 
 std::string llama_chat_format_example(const struct llama_model * model,

common/common.h

@@ -200,6 +200,7 @@ struct gpt_params {
     std::string public_path   = "";
     std::string chat_template = "";
     std::string system_prompt = "";
+    bool enable_chat_template = true;
 
     std::vector<std::string> api_keys;
 
@@ -250,6 +251,8 @@ struct gpt_params {
     std::string cvector_outfile       = "control_vector.gguf";
     std::string cvector_positive_file = "examples/cvector-generator/positive.txt";
     std::string cvector_negative_file = "examples/cvector-generator/negative.txt";
+
+    bool spm_infill = false; // suffix/prefix/middle pattern for infill
 };
 
 void gpt_params_handle_model_default(gpt_params & params);
@@ -456,4 +459,3 @@ void yaml_dump_string_multiline(FILE * stream, const char * prop_name, const cha
 void yaml_dump_non_result_info(
     FILE * stream, const gpt_params & params, const llama_context * lctx,
     const std::string & timestamp, const std::vector<int> & prompt_tokens, const char * model_desc);
-

common/json-schema-to-grammar.cpp

@@ -316,7 +316,7 @@ std::unordered_map<char, std::string> GRAMMAR_LITERAL_ESCAPES = {
 };
 
 std::unordered_set<char> NON_LITERAL_SET = {'|', '.', '(', ')', '[', ']', '{', '}', '*', '+', '?'};
-std::unordered_set<char> ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = {'[', ']', '(', ')', '|', '{', '}', '*', '+', '?'};
+std::unordered_set<char> ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = {'^', '$', '.', '[', ']', '(', ')', '|', '{', '}', '*', '+', '?'};
 
 template <typename Iterator>
 std::string join(Iterator begin, Iterator end, const std::string & separator) {
@@ -720,7 +720,7 @@ class SchemaConverter {
             }
             prop_names.push_back(prop_name);
         }
-        if (!(additional_properties.is_boolean() && !additional_properties.get<bool>())) {
+        if ((additional_properties.is_boolean() && additional_properties.get<bool>()) || additional_properties.is_object()) {
             std::string sub_name = name + (name.empty() ? "" : "-") + "additional";
             std::string value_rule =
                 additional_properties.is_object() ? visit(additional_properties, sub_name + "-value")

convert-hf-to-gguf-update.py

@@ -86,6 +86,7 @@ class TOKENIZER_TYPE(IntEnum):
     {"name": "poro-chat",      "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Poro-34B-chat", },
     {"name": "jina-v2-code",   "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-code", },
     {"name": "viking",         "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Viking-7B", }, # Also used for Viking 13B and 33B
+    {"name": "jais",           "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/core42/jais-13b", },
 ]