node-llama-cpp 3.17.1 → 3.18.1

package/llama/addon/AddonContext.h

@@ -1,4 +1,7 @@
 #pragma once
+
+#include <mutex>
+
 #include "llama.h"
 #include "napi.h"
 #include "addonGlobals.h"
@@ -53,6 +56,30 @@ class AddonContext : public Napi::ObjectWrap<AddonContext> {
         Napi::Value EnsureDraftContextIsCompatibleForSpeculative(const Napi::CallbackInfo& info);
 
         Napi::Value SetLoras(const Napi::CallbackInfo& info);
+        Napi::Value RestoreCheckpoint(const Napi::CallbackInfo& info);
+
+        static void init(Napi::Object exports);
+};
+
+class AddonContextSequenceCheckpoint : public Napi::ObjectWrap<AddonContextSequenceCheckpoint> {
+    public:
+        std::mutex dataMutex;
+        std::vector<uint8_t> data;
+        llama_seq_id sequenceId = 0;
+        std::size_t minPos = 0;
+        std::size_t maxPos = 0;
+
+        AddonContextSequenceCheckpoint(const Napi::CallbackInfo& info);
+        ~AddonContextSequenceCheckpoint();
+
+        Napi::Value Init(const Napi::CallbackInfo& info);
+        Napi::Value Dispose(const Napi::CallbackInfo& info);
+
+        void dispose();
+
+        Napi::Value GetSize(const Napi::CallbackInfo& info);
+        Napi::Value GetMinPos(const Napi::CallbackInfo& info);
+        Napi::Value GetMaxPos(const Napi::CallbackInfo& info);
 
         static void init(Napi::Object exports);
 };

package/llama/addon/addon.cpp

@@ -317,6 +317,7 @@ Napi::Object registerCallback(Napi::Env env, Napi::Object exports) {
     AddonGrammar::init(exports);
     AddonGrammarEvaluationState::init(exports);
     AddonContext::init(exports);
+    AddonContextSequenceCheckpoint::init(exports);
     AddonSampler::init(exports);
 
     llama_log_set(addonLlamaCppLogCallback, nullptr);

package/llama/binariesGithubRelease.json

@@ -1,3 +1,3 @@
 {
-    "release": "b8179"
+    "release": "b8390"
 }

package/llama/llama.cpp.info.json

@@ -1,4 +1,4 @@
 {
-    "tag": "b8179",
+    "tag": "b8390",
     "llamaCppGithubRepo": "ggml-org/llama.cpp"
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "node-llama-cpp",
-  "version": "3.17.1",
+  "version": "3.18.1",
   "description": "Run AI models locally on your machine with node.js bindings for llama.cpp. Enforce a JSON schema on the model output on the generation level",
   "main": "./dist/index.js",
   "type": "module",
@@ -145,7 +145,7 @@
     "@resvg/resvg-js": "^2.6.2",
     "@semantic-release/exec": "^7.1.0",
     "@semantic-release/github": "^12.0.6",
-    "@semantic-release/npm": "^13.1.4",
+    "@semantic-release/npm": "^13.1.5",
     "@shikijs/vitepress-twoslash": "^3.22.0",
     "@stylistic/eslint-plugin": "^5.8.0",
     "@types/async-retry": "^1.4.9",
@@ -158,8 +158,8 @@
     "@types/validate-npm-package-name": "^4.0.2",
     "@types/which": "^3.0.4",
     "@types/yargs": "^17.0.33",
-    "@vitest/coverage-v8": "^4.0.18",
-    "@vitest/ui": "^4.0.18",
+    "@vitest/coverage-v8": "^4.1.0",
+    "@vitest/ui": "^4.1.0",
     "electron": "^40.4.1",
     "eslint": "^9.39.2",
     "eslint-import-resolver-typescript": "^4.4.4",
@@ -179,14 +179,14 @@
     "typedoc-vitepress-theme": "^1.1.2",
     "typescript": "^5.9.3",
     "typescript-eslint": "^8.56.0",
-    "vite-node": "^5.3.0",
+    "vite-node": "^6.0.0",
     "vitepress": "^1.6.4",
     "vitepress-plugin-llms": "^1.11.0",
-    "vitest": "^4.0.18",
+    "vitest": "^4.1.0",
     "zx": "^8.8.5"
   },
   "dependencies": {
-    "@huggingface/jinja": "^0.5.5",
+    "@huggingface/jinja": "^0.5.6",
     "async-retry": "^1.3.3",
     "bytes": "^3.1.2",
     "chalk": "^5.6.2",
@@ -195,22 +195,22 @@
     "cross-spawn": "^7.0.6",
     "env-var": "^7.5.0",
     "filenamify": "^6.0.0",
-    "fs-extra": "^11.3.0",
+    "fs-extra": "^11.3.4",
     "ignore": "^7.0.4",
     "ipull": "^3.9.5",
     "is-unicode-supported": "^2.1.0",
     "lifecycle-utils": "^3.1.1",
     "log-symbols": "^7.0.1",
     "nanoid": "^5.1.6",
-    "node-addon-api": "^8.5.0",
+    "node-addon-api": "^8.6.0",
     "ora": "^9.3.0",
     "pretty-ms": "^9.3.0",
     "proper-lockfile": "^4.1.2",
     "semver": "^7.7.1",
-    "simple-git": "^3.32.2",
+    "simple-git": "^3.33.0",
     "slice-ansi": "^8.0.0",
     "stdout-update": "^4.0.1",
-    "strip-ansi": "^7.1.2",
+    "strip-ansi": "^7.2.0",
     "validate-npm-package-name": "^7.0.2",
     "which": "^6.0.1",
     "yargs": "^17.7.2"
@@ -224,18 +224,18 @@
     }
   },
   "optionalDependencies": {
-    "@node-llama-cpp/linux-arm64": "3.17.1",
-    "@node-llama-cpp/linux-armv7l": "3.17.1",
-    "@node-llama-cpp/linux-x64": "3.17.1",
-    "@node-llama-cpp/linux-x64-cuda": "3.17.1",
-    "@node-llama-cpp/linux-x64-cuda-ext": "3.17.1",
-    "@node-llama-cpp/linux-x64-vulkan": "3.17.1",
-    "@node-llama-cpp/mac-arm64-metal": "3.17.1",
-    "@node-llama-cpp/mac-x64": "3.17.1",
-    "@node-llama-cpp/win-arm64": "3.17.1",
-    "@node-llama-cpp/win-x64": "3.17.1",
-    "@node-llama-cpp/win-x64-cuda": "3.17.1",
-    "@node-llama-cpp/win-x64-cuda-ext": "3.17.1",
-    "@node-llama-cpp/win-x64-vulkan": "3.17.1"
+    "@node-llama-cpp/linux-arm64": "3.18.1",
+    "@node-llama-cpp/linux-armv7l": "3.18.1",
+    "@node-llama-cpp/linux-x64": "3.18.1",
+    "@node-llama-cpp/linux-x64-cuda": "3.18.1",
+    "@node-llama-cpp/linux-x64-cuda-ext": "3.18.1",
+    "@node-llama-cpp/linux-x64-vulkan": "3.18.1",
+    "@node-llama-cpp/mac-arm64-metal": "3.18.1",
+    "@node-llama-cpp/mac-x64": "3.18.1",
+    "@node-llama-cpp/win-arm64": "3.18.1",
+    "@node-llama-cpp/win-x64": "3.18.1",
+    "@node-llama-cpp/win-x64-cuda": "3.18.1",
+    "@node-llama-cpp/win-x64-cuda-ext": "3.18.1",
+    "@node-llama-cpp/win-x64-vulkan": "3.18.1"
   }
 }