update ggml

marty1885 · Nov 2, 2024 · ad22912 · ad22912
2 parents 3105185 + a6744e4
commit ad22912
Show file tree

Hide file tree

Showing 52 changed files with 3,431 additions and 2,740 deletions.
diff --git a/Makefile b/Makefile
@@ -34,6 +34,7 @@ BUILD_TARGETS = \
 	llama-save-load-state \
 	llama-server \
 	llama-simple \
+	llama-simple-chat \
 	llama-speculative \
 	llama-tokenize \
 	llama-vdot \
@@ -1287,6 +1288,11 @@ llama-simple: examples/simple/simple.cpp \
 	$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
 
+llama-simple-chat: examples/simple-chat/simple-chat.cpp \
+	$(OBJ_ALL)
+	$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
+	$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
+
 llama-tokenize: examples/tokenize/tokenize.cpp \
 	$(OBJ_ALL)
 	$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)

diff --git a/README.md b/README.md
@@ -17,7 +17,8 @@ Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others)
 
 ## Hot topics
 
-- **Hugging Face Inference Endpoints now support GGUF out of the box! https://github.com/ggerganov/llama.cpp/discussions/9669**
+- **Introducing GGUF-my-LoRA** https://github.com/ggerganov/llama.cpp/discussions/10123
+- Hugging Face Inference Endpoints now support GGUF out of the box! https://github.com/ggerganov/llama.cpp/discussions/9669
 - Hugging Face GGUF editor: [discussion](https://github.com/ggerganov/llama.cpp/discussions/9268) | [tool](https://huggingface.co/spaces/CISCai/gguf-editor)
 
 ----

diff --git a/common/arg.cpp b/common/arg.cpp
@@ -943,13 +943,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
             params.sparams.min_p = std::stof(value);
         }
     ).set_sparam());
-    add_opt(common_arg(
-        {"--tfs"}, "N",
-        string_format("tail free sampling, parameter z (default: %.1f, 1.0 = disabled)", (double)params.sparams.tfs_z),
-        [](common_params & params, const std::string & value) {
-            params.sparams.tfs_z = std::stof(value);
-        }
-    ).set_sparam());
     add_opt(common_arg(
         {"--xtc-probability"}, "N",
         string_format("xtc probability (default: %.1f, 0.0 = disabled)", (double)params.sparams.xtc_probability),
@@ -1074,7 +1067,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
     ).set_sparam());
     add_opt(common_arg(
         {"--mirostat"}, "N",
-        string_format("use Mirostat sampling.\nTop K, Nucleus, Tail Free and Locally Typical samplers are ignored if used.\n"
+        string_format("use Mirostat sampling.\nTop K, Nucleus and Locally Typical samplers are ignored if used.\n"
         "(default: %d, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0)", params.sparams.mirostat),
         [](common_params & params, int value) {
             params.sparams.mirostat = value;

diff --git a/common/common.cpp b/common/common.cpp
@@ -2090,7 +2090,6 @@ void yaml_dump_non_result_info(FILE * stream, const common_params & params, cons
     const std::vector<float> tensor_split_vector(params.tensor_split, params.tensor_split + llama_max_devices());
     yaml_dump_vector_float(stream, "tensor_split", tensor_split_vector);
 
-    fprintf(stream, "tfs: %f # default: 1.0\n", sparams.tfs_z);
     fprintf(stream, "threads: %d # default: %u\n", params.cpuparams.n_threads, std::thread::hardware_concurrency());
     fprintf(stream, "top_k: %d # default: 40\n", sparams.top_k);
     fprintf(stream, "top_p: %f # default: 0.95\n", sparams.top_p);

diff --git a/common/common.h b/common/common.h
@@ -88,7 +88,7 @@ enum common_sampler_type {
     COMMON_SAMPLER_TYPE_TOP_K       = 2,
     COMMON_SAMPLER_TYPE_TOP_P       = 3,
     COMMON_SAMPLER_TYPE_MIN_P       = 4,
-    COMMON_SAMPLER_TYPE_TFS_Z       = 5,
+  //COMMON_SAMPLER_TYPE_TFS_Z       = 5,
     COMMON_SAMPLER_TYPE_TYPICAL_P   = 6,
     COMMON_SAMPLER_TYPE_TEMPERATURE = 7,
     COMMON_SAMPLER_TYPE_XTC         = 8,
@@ -113,7 +113,6 @@ struct common_sampler_params {
     float   min_p              = 0.05f; // 0.0 = disabled
     float   xtc_probability    = 0.00f; // 0.0 = disabled
     float   xtc_threshold      = 0.10f; // > 0.5 disables XTC
-    float   tfs_z              = 1.00f; // 1.0 = disabled
     float   typ_p              = 1.00f; // typical_p, 1.0 = disabled
     float   temp               = 0.80f; // <= 0.0 to sample greedily, 0.0 to not output probabilities
     float   dynatemp_range     = 0.00f; // 0.0 = disabled
@@ -139,7 +138,6 @@ struct common_sampler_params {
     std::vector<enum common_sampler_type> samplers = {
         COMMON_SAMPLER_TYPE_DRY,
         COMMON_SAMPLER_TYPE_TOP_K,
-        COMMON_SAMPLER_TYPE_TFS_Z,
         COMMON_SAMPLER_TYPE_TYPICAL_P,
         COMMON_SAMPLER_TYPE_TOP_P,
         COMMON_SAMPLER_TYPE_MIN_P,

diff --git a/common/sampling.cpp b/common/sampling.cpp
@@ -131,11 +131,11 @@ std::string common_sampler_params::print() const {
     snprintf(result, sizeof(result),
             "\trepeat_last_n = %d, repeat_penalty = %.3f, frequency_penalty = %.3f, presence_penalty = %.3f\n"
             "\tdry_multiplier = %.3f, dry_base = %.3f, dry_allowed_length = %d, dry_penalty_last_n = %d\n"
-            "\ttop_k = %d, tfs_z = %.3f, top_p = %.3f, min_p = %.3f, xtc_probability = %.3f, xtc_threshold = %.3f, typical_p = %.3f, temp = %.3f\n"
+            "\ttop_k = %d, top_p = %.3f, min_p = %.3f, xtc_probability = %.3f, xtc_threshold = %.3f, typical_p = %.3f, temp = %.3f\n"
             "\tmirostat = %d, mirostat_lr = %.3f, mirostat_ent = %.3f",
             penalty_last_n, penalty_repeat, penalty_freq, penalty_present,
             dry_multiplier, dry_base, dry_allowed_length, dry_penalty_last_n,
-            top_k, tfs_z, top_p, min_p, xtc_probability, xtc_threshold, typ_p, temp,
+            top_k, top_p, min_p, xtc_probability, xtc_threshold, typ_p, temp,
             mirostat, mirostat_eta, mirostat_tau);
 
     return std::string(result);
@@ -199,9 +199,6 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, co
                 case COMMON_SAMPLER_TYPE_XTC:
                     llama_sampler_chain_add(result->chain, llama_sampler_init_xtc      (params.xtc_probability, params.xtc_threshold, params.min_keep, params.seed));
                     break;
-                case COMMON_SAMPLER_TYPE_TFS_Z:
-                    llama_sampler_chain_add(result->chain, llama_sampler_init_tail_free(params.tfs_z, params.min_keep));
-                    break;
                 case COMMON_SAMPLER_TYPE_TYPICAL_P:
                     llama_sampler_chain_add(result->chain, llama_sampler_init_typical  (params.typ_p, params.min_keep));
                     break;
@@ -373,7 +370,6 @@ char common_sampler_type_to_chr(enum common_sampler_type cnstr) {
     switch (cnstr) {
         case COMMON_SAMPLER_TYPE_DRY:         return 'd';
         case COMMON_SAMPLER_TYPE_TOP_K:       return 'k';
-        case COMMON_SAMPLER_TYPE_TFS_Z:       return 'f';
         case COMMON_SAMPLER_TYPE_TYPICAL_P:   return 'y';
         case COMMON_SAMPLER_TYPE_TOP_P:       return 'p';
         case COMMON_SAMPLER_TYPE_MIN_P:       return 'm';
@@ -388,7 +384,6 @@ std::string common_sampler_type_to_str(enum common_sampler_type cnstr) {
     switch (cnstr) {
         case COMMON_SAMPLER_TYPE_DRY:         return "dry";
         case COMMON_SAMPLER_TYPE_TOP_K:       return "top_k";
-        case COMMON_SAMPLER_TYPE_TFS_Z:       return "tfs_z";
         case COMMON_SAMPLER_TYPE_TYPICAL_P:   return "typ_p";
         case COMMON_SAMPLER_TYPE_TOP_P:       return "top_p";
         case COMMON_SAMPLER_TYPE_MIN_P:       return "min_p";
@@ -406,7 +401,6 @@ std::vector<common_sampler_type> common_sampler_types_from_names(const std::vect
         { "top_p",       COMMON_SAMPLER_TYPE_TOP_P },
         { "typ_p",       COMMON_SAMPLER_TYPE_TYPICAL_P },
         { "min_p",       COMMON_SAMPLER_TYPE_MIN_P },
-        { "tfs_z",       COMMON_SAMPLER_TYPE_TFS_Z },
         { "temperature", COMMON_SAMPLER_TYPE_TEMPERATURE },
         { "xtc",         COMMON_SAMPLER_TYPE_XTC },
         { "infill",      COMMON_SAMPLER_TYPE_INFILL },
@@ -423,8 +417,6 @@ std::vector<common_sampler_type> common_sampler_types_from_names(const std::vect
         { "typ-p",       COMMON_SAMPLER_TYPE_TYPICAL_P },
         { "typ",         COMMON_SAMPLER_TYPE_TYPICAL_P },
         { "min-p",       COMMON_SAMPLER_TYPE_MIN_P },
-        { "tfs-z",       COMMON_SAMPLER_TYPE_TFS_Z },
-        { "tfs",         COMMON_SAMPLER_TYPE_TFS_Z },
         { "temp",        COMMON_SAMPLER_TYPE_TEMPERATURE },
     };
 
@@ -452,7 +444,6 @@ std::vector<common_sampler_type> common_sampler_types_from_chars(const std::stri
     std::unordered_map<char, common_sampler_type> sampler_name_map = {
         { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_DRY),         COMMON_SAMPLER_TYPE_DRY },
         { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_TOP_K),       COMMON_SAMPLER_TYPE_TOP_K },
-        { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_TFS_Z),       COMMON_SAMPLER_TYPE_TFS_Z },
         { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_TYPICAL_P),   COMMON_SAMPLER_TYPE_TYPICAL_P },
         { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_TOP_P),       COMMON_SAMPLER_TYPE_TOP_P },
         { common_sampler_type_to_chr(COMMON_SAMPLER_TYPE_MIN_P),       COMMON_SAMPLER_TYPE_MIN_P },

diff --git a/convert_lora_to_gguf.py b/convert_lora_to_gguf.py
@@ -230,7 +230,7 @@ def get_base_tensor_name(lora_tensor_name: str) -> str:
 
 def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
-        description="Convert a huggingface PEFT LoRA adapter to a GGML compatible file")
+        description="Convert a Hugging Face PEFT LoRA adapter to a GGUF file")
     parser.add_argument(
         "--outfile", type=Path,
         help="path to write to; default: based on input. {ftype} will be replaced by the outtype.",
@@ -257,11 +257,11 @@ def parse_args() -> argparse.Namespace:
     )
     parser.add_argument(
         "--base", type=Path, required=True,
-        help="directory containing base model file",
+        help="directory containing Hugging Face model config files (config.json, tokenizer.json) for the base model that the adapter is based on - only config is needed, actual model weights are not required",
     )
     parser.add_argument(
         "lora_path", type=Path,
-        help="directory containing LoRA adapter file",
+        help="directory containing Hugging Face PEFT LoRA config (adapter_model.json) and weights (adapter_model.safetensors or adapter_model.bin)",
     )
 
     return parser.parse_args()

diff --git a/examples/CMakeLists.txt b/examples/CMakeLists.txt
@@ -49,6 +49,7 @@ else()
     endif()
     add_subdirectory(save-load-state)
     add_subdirectory(simple)
+    add_subdirectory(simple-chat)
     add_subdirectory(speculative)
     add_subdirectory(tokenize)
 endif()
diff --git a/examples/llama-bench/llama-bench.cpp b/examples/llama-bench/llama-bench.cpp
@@ -21,12 +21,6 @@
 #include "ggml.h"
 #include "llama.h"
 #include "common.h"
-#include "ggml-cuda.h"
-#include "ggml-sycl.h"
-
-#ifdef GGML_USE_CANN
-#include "ggml-cann.h"
-#endif
 
 #ifdef _WIN32
 #define WIN32_LEAN_AND_MEAN
@@ -82,95 +76,27 @@ static T stdev(const std::vector<T> & v) {
 }
 
 static std::string get_cpu_info() {
-    std::string id;
-#ifdef __linux__
-    FILE * f = fopen("/proc/cpuinfo", "r");
-    if (f) {
-        char buf[1024];
-        while (fgets(buf, sizeof(buf), f)) {
-            if (strncmp(buf, "model name", 10) == 0) {
-                char * p = strchr(buf, ':');
-                if (p) {
-                    p++;
-                    while (std::isspace(*p)) {
-                        p++;
-                    }
-                    while (std::isspace(p[strlen(p) - 1])) {
-                        p[strlen(p) - 1] = '\0';
-                    }
-                    id = p;
-                    break;
-                }
-            }
+    std::vector<std::string> cpu_list;
+    for (size_t i = 0; i < ggml_backend_dev_count(); i++) {
+        auto * dev = ggml_backend_dev_get(i);
+        auto dev_type = ggml_backend_dev_type(dev);
+        if (dev_type == GGML_BACKEND_DEVICE_TYPE_CPU || dev_type == GGML_BACKEND_DEVICE_TYPE_ACCEL) {
+            cpu_list.push_back(ggml_backend_dev_description(dev));
         }
-        fclose(f);
-    }
-#elif defined(_WIN32)
-    HKEY hKey;
-    if (RegOpenKeyEx(HKEY_LOCAL_MACHINE,
-                     TEXT("HARDWARE\\DESCRIPTION\\System\\CentralProcessor\\0"),
-                     0,
-                     KEY_READ,
-                     &hKey) != ERROR_SUCCESS) {
-        // fail to open registry key
-        return "";
     }
-    char cpu_brand[256];
-    DWORD cpu_brand_size = sizeof(cpu_brand);
-    if (RegQueryValueExA(hKey,
-                        TEXT("ProcessorNameString"),
-                        NULL,
-                        NULL,
-                        (LPBYTE)cpu_brand,
-                        &cpu_brand_size) == ERROR_SUCCESS) {
-        id.assign(cpu_brand, cpu_brand_size);
-        if (id.find('\0') != std::string::npos) {
-            id.resize(id.find('\0'));
-        }
-    }
-    RegCloseKey(hKey);
-#endif
-    // TODO: other platforms
-    return id;
+    return join(cpu_list, ", ");
 }
 
 static std::string get_gpu_info() {
-    std::string id;
-#ifdef GGML_USE_CUDA
-    int count = ggml_backend_cuda_get_device_count();
-    for (int i = 0; i < count; i++) {
-        char buf[128];
-        ggml_backend_cuda_get_device_description(i, buf, sizeof(buf));
-        id += buf;
-        if (i < count - 1) {
-            id += "/";
+    std::vector<std::string> gpu_list;
+    for (size_t i = 0; i < ggml_backend_dev_count(); i++) {
+        auto * dev = ggml_backend_dev_get(i);
+        auto dev_type = ggml_backend_dev_type(dev);
+        if (dev_type == GGML_BACKEND_DEVICE_TYPE_GPU) {
+            gpu_list.push_back(ggml_backend_dev_description(dev));
         }
     }
-#endif
-#ifdef GGML_USE_SYCL
-    int count = ggml_backend_sycl_get_device_count();
-    for (int i = 0; i < count; i++) {
-        char buf[128];
-        ggml_backend_sycl_get_device_description(i, buf, sizeof(buf));
-        id += buf;
-        if (i < count - 1) {
-            id += "/";
-        }
-    }
-#endif
-#ifdef GGML_USE_CANN
-    uint32_t count = ggml_backend_cann_get_device_count();
-    for (uint32_t i = 0; i < count; i++) {
-        char buf[128];
-        ggml_backend_cann_get_device_description(i, buf, sizeof(buf));
-        id += buf;
-        if (i < count - 1) {
-            id += "/";
-        }
-    }
-#endif
-    // TODO: other backends
-    return id;
+    return join(gpu_list, ", ");
 }
 
 // command line params
@@ -938,29 +864,15 @@ struct test {
     }
 
     static std::string get_backend() {
-        if (cuda) {
-            return GGML_CUDA_NAME;
-        }
-        if (vulkan) {
-            return "Vulkan";
-        }
-        if (kompute) {
-            return "Kompute";
-        }
-        if (metal) {
-            return "Metal";
-        }
-        if (sycl) {
-            return GGML_SYCL_NAME;
-        }
-        if (gpu_blas) {
-            return "GPU BLAS";
-        }
-        if (blas) {
-            return "BLAS";
+        std::vector<std::string> backends;
+        for (size_t i = 0; i < ggml_backend_reg_count(); i++) {
+            auto * reg = ggml_backend_reg_get(i);
+            std::string name = ggml_backend_reg_name(reg);
+            if (name != "CPU") {
+                backends.push_back(ggml_backend_reg_name(reg));
+            }
         }
-
-        return "CPU";
+        return backends.empty() ? "CPU" : join(backends, ",");
     }
 
     static const std::vector<std::string> & get_fields() {

diff --git a/examples/main/README.md b/examples/main/README.md
@@ -235,14 +235,6 @@ The Min-P sampling method was designed as an alternative to Top-P, and aims to e
 
 Example usage: `--min-p 0.05`
 
-### Tail-Free Sampling (TFS)
-
--   `--tfs N`: Enable tail free sampling with parameter z (default: 1.0, 1.0 = disabled).
-
-Tail-free sampling (TFS) is a text generation technique that aims to reduce the impact of less likely tokens, which may be less relevant, less coherent, or nonsensical, on the output. Similar to Top-P it tries to determine the bulk of the most likely tokens dynamically. But TFS filters out logits based on the second derivative of their probabilities. Adding tokens is stopped after the sum of the second derivatives reaches the parameter z. In short: TFS looks at how quickly the probabilities of the tokens decrease and cuts off the tail of unlikely tokens using the parameter z. Typical values for z are in the range of 0.9 to 0.95. A value of 1.0 would include all tokens and thus disables the effect of TFS.
-
-Example usage: `--tfs 0.95`
-
 ### Locally Typical Sampling
 
 -   `--typical N`: Enable locally typical sampling with parameter p (default: 1.0, 1.0 = disabled).
@@ -341,6 +333,15 @@ These options help improve the performance and memory usage of the LLaMA models.
 
 For information about 4-bit quantization, which can significantly improve performance and reduce memory usage, please refer to llama.cpp's primary [README](../../README.md#prepare-and-quantize).
 
+## LoRA (Low-Rank Adaptation) adapters
+
+-   `--lora FNAME`: Optional path to a LoRA adapter to use with scaling of 1.0. Can be mixed with `--lora-scaled` and can be repeated to use multiple adapters.
+-   `--lora-scaled FNAME`: Optional path to a LoRA adapter with user-defined scaling. Can be mixed with `--lora` and can repeated to use multiple adapters.
+
+You can add LoRA adapters using `--lora` or `--lora-scaled`. For example: `--lora my_adapter_1.gguf --lora my_adapter_2.gguf ...` or `--lora-scaled lora_task_A.gguf 0.5 --lora-scaled lora_task_B.gguf 0.5`.
+
+LoRA adapters should be in GGUF format. To convert from Hugging Face format use the `convert-lora-to-gguf.py` script. LoRA adapters are loaded separately and applied during inference - they are not merged with the main model. This means that mmap model loading is fully supported when using LoRA adapters. The old `--lora-base` flag has been removed now that merging is no longer performed.
+
 ## Additional Options
 
 These options provide extra functionality and customization when running the LLaMA models:
@@ -349,6 +350,4 @@ These options provide extra functionality and customization when running the LLa
 -   `--verbose-prompt`: Print the prompt before generating text.
 -   `-mg i, --main-gpu i`: When using multiple GPUs this option controls which GPU is used for small tensors for which the overhead of splitting the computation across all GPUs is not worthwhile. The GPU in question will use slightly more VRAM to store a scratch buffer for temporary results. By default GPU 0 is used.
 -   `-ts SPLIT, --tensor-split SPLIT`: When using multiple GPUs this option controls how large tensors should be split across all GPUs. `SPLIT` is a comma-separated list of non-negative values that assigns the proportion of data that each GPU should get in order. For example, "3,2" will assign 60% of the data to GPU 0 and 40% to GPU 1. By default the data is split in proportion to VRAM but this may not be optimal for performance.
--   `--lora FNAME`: Apply a LoRA (Low-Rank Adaptation) adapter to the model (implies --no-mmap). This allows you to adapt the pretrained model to specific tasks or domains.
--   `--lora-base FNAME`: Optional model to use as a base for the layers modified by the LoRA adapter. This flag is used in conjunction with the `--lora` flag, and specifies the base model for the adaptation.
 -   `-hfr URL --hf-repo URL`: The url to the Hugging Face model repository. Used in conjunction with `--hf-file` or `-hff`. The model is downloaded and stored in the file provided by `-m` or `--model`. If `-m` is not provided, the model is auto-stored in the path specified by the `LLAMA_CACHE` environment variable  or in an OS-specific local cache.