llama_cpp 0.2.0 → 0.2.2

data/ext/llama_cpp/src/llama.h

@@ -71,27 +71,27 @@ extern "C" {
 
     typedef void (*llama_progress_callback)(float progress, void *ctx);
 
-    struct llama_context_params {
+   struct llama_context_params {
+        int seed;                              // RNG seed, -1 for random
         int n_ctx;                             // text context
         int n_batch;                           // prompt processing batch size
         int n_gpu_layers;                      // number of layers to store in VRAM
         int main_gpu;                          // the GPU that is used for scratch and small tensors
         float tensor_split[LLAMA_MAX_DEVICES]; // how to split layers across multiple GPUs
-        int seed;                              // RNG seed, -1 for random
+        // called with a progress value between 0 and 1, pass NULL to disable
+        llama_progress_callback progress_callback;
+        // context pointer passed to the progress callback
+        void * progress_callback_user_data;
 
+        // Keep the booleans together to avoid misalignment during copy-by-value.
+        bool low_vram;   // if true, reduce VRAM usage at the cost of performance
         bool f16_kv;     // use fp16 for KV cache
         bool logits_all; // the llama_eval() call computes all logits, not just the last one
         bool vocab_only; // only load the vocabulary, no weights
         bool use_mmap;   // use mmap if possible
         bool use_mlock;  // force system to keep model in RAM
         bool embedding;  // embedding mode only
-
-        // called with a progress value between 0 and 1, pass NULL to disable
-        llama_progress_callback progress_callback;
-        // context pointer passed to the progress callback
-        void * progress_callback_user_data;
     };
-
     // model file types
     enum llama_ftype {
         LLAMA_FTYPE_ALL_F32              = 0,
@@ -220,6 +220,14 @@ extern "C" {
     LLAMA_API int llama_n_ctx  (const struct llama_context * ctx);
     LLAMA_API int llama_n_embd (const struct llama_context * ctx);
 
+    // Get the vocabulary as output parameters.
+    // Returns number of results.
+    LLAMA_API int llama_get_vocab(
+            const struct llama_context * ctx,
+                          const char * * strings,
+                                 float * scores,
+                                   int   capacity);
+
     // Token logits obtained from the last call to llama_eval()
     // The logits for the last token are stored in the last row
     // Can be mutated in order to change the probabilities of the next token
@@ -235,9 +243,9 @@ extern "C" {
     LLAMA_API const char * llama_token_to_str(const struct llama_context * ctx, llama_token token);
 
     // Special tokens
-    LLAMA_API llama_token llama_token_bos();
-    LLAMA_API llama_token llama_token_eos();
-    LLAMA_API llama_token llama_token_nl();
+    LLAMA_API llama_token llama_token_bos();  // beginning-of-sentence
+    LLAMA_API llama_token llama_token_eos();  // end-of-sentence
+    LLAMA_API llama_token llama_token_nl();   // next-line
 
     // Sampling functions
 

data/lib/llama_cpp/version.rb

@@ -3,8 +3,8 @@
 # llama_cpp.rb provides Ruby bindings for the llama.cpp.
 module LLaMACpp
   # The version of llama_cpp.rb you install.
-  VERSION = '0.2.0'
+  VERSION = '0.2.2'
 
   # The version of llama.cpp bundled with llama_cpp.rb.
-  LLAMA_CPP_VERSION = 'master-4de0334'
+  LLAMA_CPP_VERSION = 'master-7487137'
 end

data/sig/llama_cpp.rbs

@@ -70,6 +70,7 @@ module LLaMACpp
     def n_ctx: () -> Integer
     def n_embd: () -> Integer
     def n_vocab: () -> Integer
+    def vocab: (capacity: Integer) -> [Array[String], Array[Float]]
     def print_timings: () -> void
     def reset_timings: () -> void
     def token_to_str: (Integer) -> String
@@ -111,6 +112,8 @@ module LLaMACpp
     def main_gpu: () -> Integer
     def main_gpu=: (Integer) -> Integer
     def tensor_split: () -> Array[Float]
+    def low_vram: () -> bool
+    def low_vram=: (bool) -> bool
     def seed: () -> Integer
     def seed=: (Integer) -> Integer
     def use_mlock: () -> bool

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: llama_cpp
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.2
 platform: ruby
 authors:
 - yoshoku
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2023-06-11 00:00:00.000000000 Z
+date: 2023-06-23 00:00:00.000000000 Z
 dependencies: []
 description: llama_cpp.rb provides Ruby bindings for the llama.cpp.
 email:
@@ -22,6 +22,9 @@ files:
 - CODE_OF_CONDUCT.md
 - LICENSE.txt
 - README.md
+- examples/README.md
+- examples/chat.rb
+- examples/embedding.rb
 - ext/llama_cpp/extconf.rb
 - ext/llama_cpp/llama_cpp.cpp
 - ext/llama_cpp/llama_cpp.h