whispercpp 1.3.0 → 1.3.1

data/ext/{whisper.h → include/whisper.h}

@@ -2,6 +2,7 @@
 #define WHISPER_H
 
 #include "ggml.h"
+#include "ggml-cpu.h"
 
 #include <stddef.h>
 #include <stdint.h>
@@ -99,6 +100,7 @@ extern "C" {
         WHISPER_AHEADS_LARGE_V1,
         WHISPER_AHEADS_LARGE_V2,
         WHISPER_AHEADS_LARGE_V3,
+        WHISPER_AHEADS_LARGE_V3_TURBO,
     };
 
     typedef struct whisper_ahead {
@@ -113,6 +115,7 @@ extern "C" {
 
     struct whisper_context_params {
         bool  use_gpu;
+        bool  flash_attn;
         int   gpu_device;  // CUDA device
 
         // [EXPERIMENTAL] Token-level timestamps with DTW
@@ -237,6 +240,13 @@ extern "C" {
     //                     GPU, by caching compiled 'blobs' there.
     //                     Set to nullptr if not used.
     // Returns 0 on success. If OpenVINO is not enabled in build, this simply returns 1.
+    WHISPER_API int whisper_ctx_init_openvino_encoder_with_state(
+        struct whisper_context * ctx,
+          struct whisper_state * state,
+                    const char * model_path,
+                    const char * device,
+                    const char * cache_dir);
+
     WHISPER_API int whisper_ctx_init_openvino_encoder(
         struct whisper_context * ctx,
                     const char * model_path,
@@ -265,22 +275,6 @@ extern "C" {
                                int   n_samples,
                                int   n_threads);
 
-    // Convert RAW PCM audio to log mel spectrogram but applies a Phase Vocoder to speed up the audio x2.
-    // The resulting spectrogram is stored inside the default state of the provided whisper context.
-    // Returns 0 on success
-    WHISPER_API int whisper_pcm_to_mel_phase_vocoder(
-        struct whisper_context * ctx,
-                   const float * samples,
-                           int   n_samples,
-                           int   n_threads);
-
-    WHISPER_API int whisper_pcm_to_mel_phase_vocoder_with_state(
-        struct whisper_context * ctx,
-          struct whisper_state * state,
-                   const float * samples,
-                           int   n_samples,
-                           int   n_threads);
-
     // This can be used to set a custom log mel spectrogram inside the default state of the provided whisper context.
     // Use this instead of whisper_pcm_to_mel() if you want to provide your own log mel spectrogram.
     // n_mel must be 80
@@ -350,7 +344,7 @@ extern "C" {
     int whisper_token_count(struct whisper_context * ctx, const char * text);
 
     // Largest language id (i.e. number of available languages - 1)
-    WHISPER_API int whisper_lang_max_id();
+    WHISPER_API int whisper_lang_max_id(void);
 
     // Return the id of the specified language, returns -1 if not found
     // Examples:
@@ -430,6 +424,14 @@ extern "C" {
     WHISPER_API whisper_token whisper_token_transcribe(struct whisper_context * ctx);
 
     // Performance information from the default state.
+    struct whisper_timings {
+        float sample_ms;
+        float encode_ms;
+        float decode_ms;
+        float batchd_ms;
+        float prompt_ms;
+    };
+    WHISPER_API struct whisper_timings * whisper_get_timings(struct whisper_context * ctx);
     WHISPER_API void whisper_print_timings(struct whisper_context * ctx);
     WHISPER_API void whisper_reset_timings(struct whisper_context * ctx);
 
@@ -498,7 +500,6 @@ extern "C" {
 
         // [EXPERIMENTAL] speed-up techniques
         // note: these can significantly reduce the quality of the output
-        bool speed_up;          // speed-up the audio by 2x using Phase Vocoder
         bool debug_mode;        // enable debug_mode provides extra info (eg. Dump log_mel)
         int  audio_ctx;         // overwrite the audio context size (0 = use default)
 
@@ -533,7 +534,7 @@ extern "C" {
         float temperature_inc;
         float entropy_thold;    // similar to OpenAI's "compression_ratio_threshold"
         float logprob_thold;
-        float no_speech_thold;  // TODO: not implemented
+        float no_speech_thold;
 
         struct {
             int best_of;    // ref: https://github.com/openai/whisper/blob/f82bc59f5ea234d4b97fb2860842ed38519f7e65/whisper/transcribe.py#L264
@@ -572,10 +573,10 @@ extern "C" {
     };
 
     // NOTE: this function allocates memory, and it is the responsibility of the caller to free the pointer - see whisper_free_context_params & whisper_free_params()
-    WHISPER_API struct whisper_context_params * whisper_context_default_params_by_ref();
-    WHISPER_API struct whisper_context_params whisper_context_default_params(void);
+    WHISPER_API struct whisper_context_params * whisper_context_default_params_by_ref(void);
+    WHISPER_API struct whisper_context_params   whisper_context_default_params       (void);
     WHISPER_API struct whisper_full_params * whisper_full_default_params_by_ref(enum whisper_sampling_strategy strategy);
-    WHISPER_API struct whisper_full_params whisper_full_default_params(enum whisper_sampling_strategy strategy);
+    WHISPER_API struct whisper_full_params   whisper_full_default_params       (enum whisper_sampling_strategy strategy);
 
     // Run the entire model: PCM -> log mel spectrogram -> encoder -> decoder -> text
     // Not thread safe for same context

data/ext/metal-embed.mk

@@ -0,0 +1,17 @@
+ggml/src/ggml-metal/ggml-metal-embed.o: \
+	ggml/src/ggml-metal/ggml-metal.metal \
+	ggml/src/ggml-metal/ggml-metal-impl.h \
+	ggml/src/ggml-common.h
+	@echo "Embedding Metal library"
+	@sed -e '/__embed_ggml-common.h__/r      ggml/src/ggml-common.h'                -e '/__embed_ggml-common.h__/d'      < ggml/src/ggml-metal/ggml-metal.metal           > ggml/src/ggml-metal/ggml-metal-embed.metal.tmp
+	@sed -e '/#include "ggml-metal-impl.h"/r ggml/src/ggml-metal/ggml-metal-impl.h' -e '/#include "ggml-metal-impl.h"/d' < ggml/src/ggml-metal/ggml-metal-embed.metal.tmp > ggml/src/ggml-metal/ggml-metal-embed.metal
+	$(eval TEMP_ASSEMBLY=$(shell mktemp -d))
+	@echo ".section __DATA, __ggml_metallib"                       >  $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".globl _ggml_metallib_start"                            >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo "_ggml_metallib_start:"                                  >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".incbin \"ggml/src/ggml-metal/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo ".globl _ggml_metallib_end"                              >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	@echo "_ggml_metallib_end:"                                    >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
+	$(CC) $(CFLAGS) -c $(TEMP_ASSEMBLY)/ggml-metal-embed.s -o $@
+	@rm -f ${TEMP_ASSEMBLY}/ggml-metal-embed.s
+	@rmdir ${TEMP_ASSEMBLY}

data/ext/metal.mk

@@ -0,0 +1,6 @@
+ggml/src/ggml-metal/ggml-metal.o: \
+	ggml/src/ggml-metal/ggml-metal.m \
+	ggml/src/ggml-metal/ggml-metal-impl.h \
+	ggml/include/ggml-metal.h \
+	ggml/include/ggml.h
+	$(CC) $(CFLAGS) -c $< -o $@