whisper.rn 0.3.9 → 0.4.0-rc.0

package/cpp/whisper.h

@@ -334,6 +334,11 @@ extern "C" {
     // If it returns false, the computation is aborted
     typedef bool (*whisper_encoder_begin_callback)(struct whisper_context * ctx, struct whisper_state * state, void * user_data);
 
+    // Abort callback
+    // If not NULL, called before ggml computation
+    // If it returns true, the computation is aborted
+    typedef bool (*whisper_abort_callback)(void * user_data);
+
     // Logits filter callback
     // Can be used to modify the logits before sampling
     // If not NULL, called after applying temperature to logits
@@ -428,6 +433,10 @@ extern "C" {
         whisper_encoder_begin_callback encoder_begin_callback;
         void * encoder_begin_callback_user_data;
 
+        // called each time before ggml computation starts
+        whisper_abort_callback abort_callback;
+        void * abort_callback_user_data;
+
         // called by each decoder to filter obtained logits
         whisper_logits_filter_callback logits_filter_callback;
         void * logits_filter_callback_user_data;
@@ -485,6 +494,7 @@ extern "C" {
 
     // Get whether the next segment is predicted as a speaker turn
     WHISPER_API bool whisper_full_get_segment_speaker_turn_next(struct whisper_context * ctx, int i_segment);
+    WHISPER_API bool whisper_full_get_segment_speaker_turn_next_from_state(struct whisper_state * state, int i_segment);
 
     // Get the text of the specified segment
     WHISPER_API const char * whisper_full_get_segment_text           (struct whisper_context * ctx, int i_segment);

package/ios/RNWhisperContext.mm

@@ -349,7 +349,7 @@ struct rnwhisper_segments_callback_data {
             params.new_segment_callback = [](struct whisper_context * ctx, struct whisper_state * /*state*/, int n_new, void * user_data) {
                 struct rnwhisper_segments_callback_data *data = (struct rnwhisper_segments_callback_data *)user_data;
                 data->total_n_new += n_new;
-                
+
                 NSString *text = @"";
                 NSMutableArray *segments = [[NSMutableArray alloc] init];
                 for (int i = data->total_n_new - n_new; i < data->total_n_new; i++) {
@@ -451,7 +451,7 @@ struct rnwhisper_segments_callback_data {
     if (options[@"maxContext"] != nil) {
         params.n_max_text_ctx = [options[@"maxContext"] intValue];
     }
-    
+
     if (options[@"offset"] != nil) {
         params.offset_ms = [options[@"offset"] intValue];
     }
@@ -467,16 +467,22 @@ struct rnwhisper_segments_callback_data {
     if (options[@"temperatureInc"] != nil) {
         params.temperature_inc = [options[@"temperature_inc"] floatValue];
     }
-    
+
     if (options[@"prompt"] != nil) {
         params.initial_prompt = [options[@"prompt"] UTF8String];
     }
 
+    // abort handler
     params.encoder_begin_callback = [](struct whisper_context * /*ctx*/, struct whisper_state * /*state*/, void * user_data) {
         bool is_aborted = *(bool*)user_data;
         return !is_aborted;
     };
     params.encoder_begin_callback_user_data = rn_whisper_assign_abort_map(jobId);
+    params.abort_callback = [](void * user_data) {
+        bool is_aborted = *(bool*)user_data;
+        return is_aborted;
+    };
+    params.abort_callback_user_data = rn_whisper_assign_abort_map(jobId);
 
     return params;
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "whisper.rn",
-  "version": "0.3.9",
+  "version": "0.4.0-rc.0",
   "description": "React Native binding of whisper.cpp",
   "main": "lib/commonjs/index",
   "module": "lib/module/index",

package/whisper-rn.podspec

@@ -1,7 +1,7 @@
 require "json"
 
 package = JSON.parse(File.read(File.join(__dir__, "package.json")))
-base_ld_flags = "-framework Accelerate"
+base_ld_flags = "-framework Accelerate -framework Foundation -framework Metal -framework MetalKit"
 base_compiler_flags = "-DWSP_GGML_USE_ACCELERATE -Wno-shorten-64-to-32"
 folly_compiler_flags = "-DFOLLY_NO_CONFIG -DFOLLY_MOBILE=1 -DFOLLY_USE_LIBCPP=1 -Wno-comma"
 
@@ -16,6 +16,11 @@ if ENV['RNWHISPER_DISABLE_COREML'] != '1' then
   base_compiler_flags += " -DWHISPER_USE_COREML -DWHISPER_COREML_ALLOW_FALLBACK"
 end
 
+# TODO: Enable Metal by default when we have use_gpu param
+if ENV["RNWHISPER_ENABLE_METAL"] == "1" then
+  base_compiler_flags += " -DWSP_GGML_USE_METAL" # -DWSP_GGML_METAL_NDEBUG
+end
+
 Pod::Spec.new do |s|
   s.name         = "whisper-rn"
   s.version      = package["version"]
@@ -28,6 +33,7 @@ Pod::Spec.new do |s|
   s.source       = { :git => "https://github.com/mybigday/whisper.rn.git", :tag => "#{s.version}" }
 
   s.source_files = "ios/**/*.{h,m,mm}", "cpp/**/*.{h,cpp,c,m,mm}"
+  s.resources = "cpp/**/*.{metal}"
 
   s.dependency "React-Core"
 
@@ -45,7 +51,7 @@ Pod::Spec.new do |s|
     s.pod_target_xcconfig = {
       "CLANG_CXX_LANGUAGE_STANDARD" => "c++17",
       "HEADER_SEARCH_PATHS" => "\"$(PODS_ROOT)/boost\"",
-      "OTHER_LDFLAGS" => "-framework Accelerate",
+      "OTHER_LDFLAGS" => base_ld_flags,
       "OTHER_CFLAGS" => base_optimizer_flags,
       "OTHER_CPLUSPLUSFLAGS" => new_arch_cpp_flags + " " + base_optimizer_flags
     }