Merge branch 'master' into xsn/llama_batch_remove_compat

README.md

@@ -31,7 +31,7 @@ variety of hardware - locally and in the cloud.
 - Apple silicon is a first-class citizen - optimized via ARM NEON, Accelerate and Metal frameworks
 - AVX, AVX2 and AVX512 support for x86 architectures
 - 1.5-bit, 2-bit, 3-bit, 4-bit, 5-bit, 6-bit, and 8-bit integer quantization for faster inference and reduced memory use
-- Custom CUDA kernels for running LLMs on NVIDIA GPUs (support for AMD GPUs via HIP)
+- Custom CUDA kernels for running LLMs on NVIDIA GPUs (support for AMD GPUs via HIP and Moore Threads MTT GPUs via MUSA)
 - Vulkan and SYCL backend support
 - CPU+GPU hybrid inference to partially accelerate models larger than the total VRAM capacity
 
@@ -413,7 +413,7 @@ Please refer to [Build llama.cpp locally](./docs/build.md)
 | [BLAS](./docs/build.md#blas-build) | All |
 | [BLIS](./docs/backend/BLIS.md) | All |
 | [SYCL](./docs/backend/SYCL.md) | Intel and Nvidia GPU |
-| [MUSA](./docs/build.md#musa) | Moore Threads GPU |
+| [MUSA](./docs/build.md#musa) | Moore Threads MTT GPU |
 | [CUDA](./docs/build.md#cuda) | Nvidia GPU |
 | [hipBLAS](./docs/build.md#hipblas) | AMD GPU |
 | [Vulkan](./docs/build.md#vulkan) | GPU |

common/common.cpp

@@ -12,6 +12,7 @@
 
 #include <algorithm>
 #include <cinttypes>
+#include <climits>
 #include <cmath>
 #include <codecvt>
 #include <cstdarg>
@@ -23,10 +24,10 @@
 #include <regex>
 #include <sstream>
 #include <string>
+#include <thread>
 #include <unordered_map>
 #include <unordered_set>
 #include <vector>
-#include <thread>
 
 #if defined(__APPLE__) && defined(__MACH__)
 #include <sys/types.h>
@@ -400,6 +401,21 @@ std::string common_params_get_system_info(const common_params & params) {
 // String utils
 //
 
+std::string string_format(const char * fmt, ...) {
+    va_list ap;
+    va_list ap2;
+    va_start(ap, fmt);
+    va_copy(ap2, ap);
+    int size = vsnprintf(NULL, 0, fmt, ap);
+    GGML_ASSERT(size >= 0 && size < INT_MAX); // NOLINT
+    std::vector<char> buf(size + 1);
+    int size2 = vsnprintf(buf.data(), size + 1, fmt, ap2);
+    GGML_ASSERT(size2 == size);
+    va_end(ap2);
+    va_end(ap);
+    return std::string(buf.data(), size);
+}
+
 std::vector<std::string> string_split(std::string input, char separator) {
     std::vector<std::string> parts;
     size_t separator_pos = input.find(separator);

common/common.h

@@ -282,7 +282,6 @@ struct common_params {
     std::string hostname      = "127.0.0.1";
     std::string public_path   = "";                                                                         // NOLINT
     std::string chat_template = "";                                                                         // NOLINT
-    std::string system_prompt = "";                                                                         // NOLINT
     bool enable_chat_template = true;
 
     std::vector<std::string> api_keys;
@@ -352,15 +351,28 @@ void common_init();
 
 std::string common_params_get_system_info(const common_params & params);
 
-bool parse_cpu_range(const std::string& range, bool(&boolmask)[GGML_MAX_N_THREADS]);
-bool parse_cpu_mask(const std::string& mask, bool(&boolmask)[GGML_MAX_N_THREADS]);
-void postprocess_cpu_params(cpu_params& cpuparams, const cpu_params* role_model = nullptr);
+bool parse_cpu_range(const std::string & range, bool(&boolmask)[GGML_MAX_N_THREADS]);
+bool parse_cpu_mask(const std::string & mask, bool(&boolmask)[GGML_MAX_N_THREADS]);
+void postprocess_cpu_params(cpu_params & cpuparams, const cpu_params * role_model = nullptr);
 bool set_process_priority(enum ggml_sched_priority prio);
 
 //
 // String utils
 //
 
+#ifdef __GNUC__
+#ifdef __MINGW32__
+#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
+#else
+#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
+#endif
+#else
+#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...)
+#endif
+
+LLAMA_COMMON_ATTRIBUTE_FORMAT(1, 2)
+std::string string_format(const char * fmt, ...);
+
 std::vector<std::string> string_split(std::string input, char separator);
 
 std::string string_strip(const std::string & str);

docs/build.md

@@ -198,6 +198,8 @@ The following compilation options are also available to tweak performance:
 
 ### MUSA
 
+This provides GPU acceleration using the MUSA cores of your Moore Threads MTT GPU. Make sure to have the MUSA SDK installed. You can download it from here: [MUSA SDK](https://developer.mthreads.com/sdk/download/musa).
+
 - Using `make`:
   ```bash
   make GGML_MUSA=1
@@ -209,6 +211,12 @@ The following compilation options are also available to tweak performance:
   cmake --build build --config Release
   ```
 
+The environment variable [`MUSA_VISIBLE_DEVICES`](https://docs.mthreads.com/musa-sdk/musa-sdk-doc-online/programming_guide/Z%E9%99%84%E5%BD%95/) can be used to specify which GPU(s) will be used.
+
+The environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` can be used to enable unified memory in Linux. This allows swapping to system RAM instead of crashing when the GPU VRAM is exhausted.
+
+Most of the compilation options available for CUDA should also be available for MUSA, though they haven't been thoroughly tested yet.
+
 ### hipBLAS
 
 This provides BLAS acceleration on HIP-supported AMD GPUs.

examples/infill/infill.cpp

@@ -205,11 +205,11 @@ int main(int argc, char ** argv) {
     std::vector<llama_token> inp_pfx = common_tokenize(ctx, params.input_prefix, false);
     std::vector<llama_token> inp_sfx = common_tokenize(ctx, params.input_suffix, false);
 
-    GGML_ASSERT(llama_token_prefix(model) >= 0);
-    GGML_ASSERT(llama_token_suffix(model) >= 0);
+    GGML_ASSERT(llama_token_fim_pre(model) >= 0);
+    GGML_ASSERT(llama_token_fim_suf(model) >= 0);
 
-    inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model));
-    inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model));
+    inp_pfx.insert(inp_pfx.begin(), llama_token_fim_pre(model));
+    inp_sfx.insert(inp_sfx.begin(), llama_token_fim_suf(model));
 
     embd_inp = params.spm_infill ? inp_sfx : inp_pfx;
     embd_end = params.spm_infill ? inp_pfx : inp_sfx;
@@ -218,7 +218,7 @@ int main(int argc, char ** argv) {
     }
     embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end());
 
-    const llama_token middle_token = llama_token_middle(model);
+    const llama_token middle_token = llama_token_fim_mid(model);
     if (middle_token >= 0) {
         embd_inp.push_back(middle_token);
     }
@@ -508,8 +508,8 @@ int main(int argc, char ** argv) {
                 std::vector<llama_token> inp_pfx = common_tokenize(ctx, params.input_prefix, false);
                 std::vector<llama_token> inp_sfx = common_tokenize(ctx, params.input_suffix, false);
 
-                inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model));
-                inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model));
+                inp_pfx.insert(inp_pfx.begin(), llama_token_fim_pre(model));
+                inp_sfx.insert(inp_sfx.begin(), llama_token_fim_suf(model));
 
                 embd_inp = params.spm_infill ? inp_sfx : inp_pfx;
                 embd_end = params.spm_infill ? inp_pfx : inp_sfx;

examples/server/README.md

@@ -60,8 +60,6 @@ The project is under active development, and we are [looking for feedback and co
 | `--yarn-attn-factor N` | YaRN: scale sqrt(t) or attention magnitude (default: 1.0)<br/>(env: LLAMA_ARG_YARN_ATTN_FACTOR) |
 | `--yarn-beta-slow N` | YaRN: high correction dim or alpha (default: 1.0)<br/>(env: LLAMA_ARG_YARN_BETA_SLOW) |
 | `--yarn-beta-fast N` | YaRN: low correction dim or beta (default: 32.0)<br/>(env: LLAMA_ARG_YARN_BETA_FAST) |
-| `-gan, --grp-attn-n N` | group-attention factor (default: 1)<br/>(env: LLAMA_ARG_GRP_ATTN_N) |
-| `-gaw, --grp-attn-w N` | group-attention width (default: 512.0)<br/>(env: LLAMA_ARG_GRP_ATTN_W) |
 | `-dkvc, --dump-kv-cache` | verbose print of the KV cache |
 | `-nkvo, --no-kv-offload` | disable KV offload<br/>(env: LLAMA_ARG_NO_KV_OFFLOAD) |
 | `-ctk, --cache-type-k TYPE` | KV cache data type for K (default: f16)<br/>(env: LLAMA_ARG_CACHE_TYPE_K) |
@@ -149,7 +147,6 @@ The project is under active development, and we are [looking for feedback and co
 | `--ssl-cert-file FNAME` | path to file a PEM-encoded SSL certificate<br/>(env: LLAMA_ARG_SSL_CERT_FILE) |
 | `-to, --timeout N` | server read/write timeout in seconds (default: 600)<br/>(env: LLAMA_ARG_TIMEOUT) |
 | `--threads-http N` | number of threads used to process HTTP requests (default: -1)<br/>(env: LLAMA_ARG_THREADS_HTTP) |
-| `-spf, --system-prompt-file FNAME` | set a file to load a system prompt (initial prompt of all slots), this is useful for chat applications |
 | `--metrics` | enable prometheus compatible metrics endpoint (default: disabled)<br/>(env: LLAMA_ARG_ENDPOINT_METRICS) |
 | `--slots` | enable slots monitoring endpoint (default: disabled)<br/>(env: LLAMA_ARG_ENDPOINT_SLOTS) |
 | `--props` | enable changing global properties via POST /props (default: disabled)<br/>(env: LLAMA_ARG_ENDPOINT_PROPS) |
@@ -320,7 +317,6 @@ node index.js
 
       - The prompt is a string or an array with the first element given as a string
       - The model's `tokenizer.ggml.add_bos_token` metadata is `true`
-      - The system prompt is empty
 
     `temperature`: Adjust the randomness of the generated text. Default: `0.8`
 
@@ -378,6 +374,8 @@ node index.js
 
     `min_keep`: If greater than 0, force samplers to return N possible tokens at minimum. Default: `0`
 
+    `t_max_predict_ms`: Set a time limit in milliseconds for the prediction (a.k.a. text-generation) phase. The timeout will trigger if the generation takes more than the specified time (measured since the first token was generated) and if a new-line character has already been generated. Useful for FIM applications. Default: `0`, which is disabled.
+
     `image_data`: An array of objects to hold base64-encoded image `data` and its `id`s to be reference in `prompt`. You can determine the place of the image in the prompt as in the following: `USER:[img-12]Describe the image in detail.\nASSISTANT:`. In this case, `[img-12]` will be replaced by the embeddings of the image with id `12` in the following `image_data` array: `{..., "image_data": [{"data": "<BASE64_STRING>", "id": 12}]}`. Use `image_data` only with multimodal models, e.g., LLaVA.
 
     `id_slot`: Assign the completion task to an specific slot. If is -1 the task will be assigned to a Idle slot.  Default: `-1`
@@ -526,7 +524,7 @@ Takes a prefix and a suffix and returns the predicted completion as stream.
 - `input_prefix`: Set the prefix of the code to infill.
 - `input_suffix`: Set the suffix of the code to infill.
 
-It also accepts all the options of `/completion` except `stream` and `prompt`.
+It also accepts all the options of `/completion`.
 
 ### **GET** `/props`: Get server global properties.
 
@@ -536,14 +534,12 @@ This endpoint is public (no API key check). By default, it is read-only. To make
 
 ```json
 {
-  "system_prompt": "",
   "default_generation_settings": { ... },
   "total_slots": 1,
   "chat_template": ""
 }
 ```
 
-- `system_prompt` - the system prompt (initial prompt of all slots). Please note that this does not take into account the chat template. It will append the prompt at the beginning of formatted prompt.
 - `default_generation_settings` - the default generation settings for the `/completion` endpoint, which has the same fields as the `generation_settings` response object from the `/completion` endpoint.
 - `total_slots` - the total number of slots for process requests (defined by `--parallel` option)
 - `chat_template` - the model's original Jinja2 prompt template
@@ -554,7 +550,7 @@ To use this endpoint with POST method, you need to start server with `--props`
 
 *Options:*
 
-- `system_prompt`: Change the system prompt (initial prompt of all slots). Please note that this does not take into account the chat template. It will append the prompt at the beginning of formatted prompt.
+- None yet
 
 ### POST `/v1/chat/completions`: OpenAI-compatible Chat Completions API