speech_recognition/esp-sr
1
0
Fork 0
mirror of https://github.com/espressif/esp-sr.git synced 2025-09-15 15:28:44 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

feat: Update API for release v2.0

Browse Source
This commit is contained in:
sxy 2022-07-08 19:25:49 +08:00
parent f76da537fd
commit ac47452d21
21 changed files with 2101 additions and 584 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

592
Kconfig.projbuild
View File

File diff suppressed because it is too large Load Diff

22
include/esp32s3/dl_lib_convq8_queue.h
View File

@ -246,6 +246,28 @@ void dl_dilation_layerq8_mc_steps(dl_convq8_queue_t **in, dl_convq8_queue_t **ou
void dl_convq8_queue_mc_bzero(dl_convq8_queue_t **cqm, int nch); void dl_convq8_queue_mc_bzero(dl_convq8_queue_t **cqm, int nch);
dl_convq8_queue_t *dl_convq8_queue_alloc_from_psram(int n, int c);
qtp_t *dl_dilation_layerq16_8(dl_convq_queue_t *in, dl_convq8_queue_t *out, int rate, int size,
dl_matrix2dq_t* filter_kernel, dl_matrix2dq_t* filter_bias,
dl_matrix2dq_t* gate_kernel, dl_matrix2dq_t* gate_bias, int prenum);
qtp_t *dl_dilation_layerq8(dl_convq8_queue_t *in, dl_convq8_queue_t *out, int rate, int size,
dl_matrix2dq8_t* filter_kernel, dl_matrix2dq_t* filter_bias,
dl_matrix2dq8_t* gate_kernel, dl_matrix2dq_t* gate_bias, int prenum);
dl_matrix2dq8_t *dl_convq8_lstm_layer(const dl_convq8_queue_t *in, dl_convq8_queue_t *out, dl_matrix2dq8_t *state_c,
dl_matrix2dq8_t *state_h, const dl_matrix2dq8_t *in_weight, const dl_matrix2dq8_t *h_weight,
const dl_matrix2dq_t *bias, int prenum);
qtp_t *dl_atrous_conv1dq8_16_s3(dl_convq8_queue_t *in, dl_convq_queue_t *out, int rate, int size,
dl_matrix2dq8_t* kernel, dl_matrix2dq_t* bias, int prenum);
void print_convq8(dl_convq8_queue_t *cq, int offset); void print_convq8(dl_convq8_queue_t *cq, int offset);
void print_convq(dl_convq_queue_t *cq, int offset); void print_convq(dl_convq_queue_t *cq, int offset);
void lstmq8_free(void);
#endif #endif

17
include/esp32s3/esp_afe_config.h
View File

@ -39,16 +39,17 @@ typedef struct {
bool vad_init; bool vad_init;
bool wakenet_init; bool wakenet_init;
bool voice_communication_init; bool voice_communication_init;
bool voip_agc_init; // AGC swich for voice communication
int voip_agc_gain; // AGC gain(dB) for voice communication
vad_mode_t vad_mode; // The value can be: VAD_MODE_0, VAD_MODE_1, VAD_MODE_2, VAD_MODE_3, VAD_MODE_4 vad_mode_t vad_mode; // The value can be: VAD_MODE_0, VAD_MODE_1, VAD_MODE_2, VAD_MODE_3, VAD_MODE_4
esp_wn_iface_t *wakenet_model; char *wakenet_model_name; // The model name of wakenet
void *wakenet_coeff;
det_mode_t wakenet_mode; det_mode_t wakenet_mode;
afe_sr_mode_t afe_mode; afe_sr_mode_t afe_mode;
int afe_perferred_core; int afe_perferred_core;
int afe_perferred_priority; int afe_perferred_priority;
int afe_ringbuf_size; int afe_ringbuf_size;
afe_memory_alloc_mode_t memory_alloc_mode; afe_memory_alloc_mode_t memory_alloc_mode;
afe_mn_peak_agc_mode_t agc_mode; afe_mn_peak_agc_mode_t agc_mode; // The agc mode for ASR
afe_pcm_config_t pcm_config; // Config the channel num of original data which is fed to the afe feed function. afe_pcm_config_t pcm_config; // Config the channel num of original data which is fed to the afe feed function.
} afe_config_t; } afe_config_t;
@ -60,9 +61,10 @@ typedef struct {
.vad_init = true, \ .vad_init = true, \
.wakenet_init = true, \ .wakenet_init = true, \
.voice_communication_init = false, \ .voice_communication_init = false, \
.voip_agc_init = false, \
.voip_agc_gain = 15, \
.vad_mode = VAD_MODE_3, \ .vad_mode = VAD_MODE_3, \
.wakenet_model = &WAKENET_MODEL, \ .wakenet_model_name = NULL, \
.wakenet_coeff = &WAKENET_COEFF, \
.wakenet_mode = DET_MODE_90, \ .wakenet_mode = DET_MODE_90, \
.afe_mode = SR_MODE_HIGH_PERF, \ .afe_mode = SR_MODE_HIGH_PERF, \
.afe_perferred_core = 0, \ .afe_perferred_core = 0, \
@ -81,9 +83,10 @@ typedef struct {
.vad_init = true, \ .vad_init = true, \
.wakenet_init = true, \ .wakenet_init = true, \
.voice_communication_init = false, \ .voice_communication_init = false, \
.voip_agc_init = false, \
.voip_agc_gain = 15, \
.vad_mode = VAD_MODE_3, \ .vad_mode = VAD_MODE_3, \
.wakenet_model = (esp_wn_iface_t *)&WAKENET_MODEL, \ .wakenet_model_name = NULL, \
.wakenet_coeff = (void *)&WAKENET_COEFF, \
.wakenet_mode = DET_MODE_2CH_90, \ .wakenet_mode = DET_MODE_2CH_90, \
.afe_mode = SR_MODE_LOW_COST, \ .afe_mode = SR_MODE_LOW_COST, \
.afe_perferred_core = 0, \ .afe_perferred_core = 0, \

68
include/esp32s3/esp_afe_sr_iface.h
View File

@ -1,10 +1,6 @@
#pragma once #pragma once
#include "stdint.h" #include "stdint.h"
#if CONFIG_AFE_INTERFACE_V1
#include "esp_afe_config.h" #include "esp_afe_config.h"
#else
#include "esp_afe_config.h"
#endif
//AFE: Audio Front-End //AFE: Audio Front-End
//SR: Speech Recognition //SR: Speech Recognition
@ -13,24 +9,30 @@
//Opaque AFE_SR data container //Opaque AFE_SR data container
typedef struct esp_afe_sr_data_t esp_afe_sr_data_t; typedef struct esp_afe_sr_data_t esp_afe_sr_data_t;
// the output state of fetch function /**
typedef enum { * @brief The state of vad
AFE_FETCH_ERROR = -3, // fetch empty data, retry it */
AFE_FETCH_CHANNEL_VERIFIED = -2, // wwe state: output channel is verified typedef enum
AFE_FETCH_NOISE = -1, // vad state: noise or silence {
AFE_FETCH_SPEECH = 0, // vad state: speech AFE_VAD_SILENCE = 0, // noise or silence
AFE_FETCH_WWE_DETECTED = 1 // wwe state: wake word is detected AFE_VAD_SPEECH // speech
} afe_fetch_mode_t; } afe_vad_state_t;
/** /**
* @brief Function to initialze a AFE_SR instance with a specified mode * @brief The result of fetch function
*
* @param mode The mode of AFE_SR
* @param perferred_core The perferred core to be pinned.
* If all task in AFE_SR can not run in real time by only one core, the another core would be used.
* @returns Handle to the AFE_SR data
*/ */
typedef esp_afe_sr_data_t* (*esp_afe_sr_iface_op_create_t)(afe_sr_mode_t mode, int perferred_cor); typedef struct afe_fetch_result_t
{
int16_t *data; // the data of audio.
int data_size; // the size of data. The unit is byte.
int wakeup_state; // the value is afe_wakeup_state_t
int wake_word_index; // if the wake word is detected. It will store the wake word index which start from 1.
int vad_state; // the value is afe_vad_state_t
int trigger_channel_id; // the channel index of output
int wake_word_length; // the length of wake word. It's unit is the number of samples.
int ret_value; // the return state of fetch function
void* reserved; // reserved for future use
} afe_fetch_result_t;
/** /**
* @brief Function to initialze a AFE_SR instance * @brief Function to initialze a AFE_SR instance
@ -54,37 +56,37 @@ typedef int (*esp_afe_sr_iface_op_get_samp_chunksize_t)(esp_afe_sr_data_t *afe);
/** /**
* @brief Get the total channel number which be config * @brief Get the total channel number which be config
* *
* @param afe The AFE_SR object to query * @param afe The AFE_SR object to query
* @return The amount of total channels * @return The amount of total channels
*/ */
typedef int (*esp_afe_sr_iface_op_get_total_channel_num_t)(esp_afe_sr_data_t *afe); typedef int (*esp_afe_sr_iface_op_get_total_channel_num_t)(esp_afe_sr_data_t *afe);
/** /**
* @brief Get the mic channel number which be config * @brief Get the mic channel number which be config
* *
* @param afe The AFE_SR object to query * @param afe The AFE_SR object to query
* @return The amount of mic channels * @return The amount of mic channels
*/ */
typedef int (*esp_afe_sr_iface_op_get_channel_num_t)(esp_afe_sr_data_t *afe); typedef int (*esp_afe_sr_iface_op_get_channel_num_t)(esp_afe_sr_data_t *afe);
/** /**
* @brief Get the sample rate of the samples to feed to the function * @brief Get the sample rate of the samples to feed to the function
* *
* @param afe The AFE_SR object to query * @param afe The AFE_SR object to query
* @return The sample rate, in hz * @return The sample rate, in hz
*/ */
typedef int (*esp_afe_sr_iface_op_get_samp_rate_t)(esp_afe_sr_data_t *afe); typedef int (*esp_afe_sr_iface_op_get_samp_rate_t)(esp_afe_sr_data_t *afe);
/** /**
* @brief Feed samples of an audio stream to the AFE_SR * @brief Feed samples of an audio stream to the AFE_SR
* *
* @Warning The input data should be arranged in the format of [CH0_0, CH1_0, ..., CHN_0, CH0_1, CH1_1, ..., CHN_1, ...]. * @Warning The input data should be arranged in the format of channel interleaving.
* The last channel is reference signal or far-end signal. * The last channel is reference signal if it has reference data.
* *
* @param afe The AFE_SR object to queryq * @param afe The AFE_SR object to query
* *
* @param in The input microphone signal, only support signed 16-bit @ 16 KHZ. The frame size can be queried by the * @param in The input microphone signal, only support signed 16-bit @ 16 KHZ. The frame size can be queried by the
* `get_samp_chunksize`. The channel number can be queried `get_channel_num`. * `get_feed_chunksize`.
* @return The size of input * @return The size of input
*/ */
typedef int (*esp_afe_sr_iface_op_feed_t)(esp_afe_sr_data_t *afe, const int16_t* in); typedef int (*esp_afe_sr_iface_op_feed_t)(esp_afe_sr_data_t *afe, const int16_t* in);
@ -95,10 +97,9 @@ typedef int (*esp_afe_sr_iface_op_feed_t)(esp_afe_sr_data_t *afe, const int16_t*
* @Warning The output is single channel data, no matter how many channels the input is. * @Warning The output is single channel data, no matter how many channels the input is.
* *
* @param afe The AFE_SR object to query * @param afe The AFE_SR object to query
* @param out The output enhanced signal. The frame size can be queried by the `get_samp_chunksize`. * @return The result of output, please refer to the definition of `afe_fetch_result_t`. (The frame size of output audio can be queried by the `get_fetch_chunksize`.)
* @return The state of output, please refer to the definition of `afe_fetch_mode_t`
*/ */
typedef afe_fetch_mode_t (*esp_afe_sr_iface_op_fetch_t)(esp_afe_sr_data_t *afe, int16_t* out); typedef afe_fetch_result_t* (*esp_afe_sr_iface_op_fetch_t)(esp_afe_sr_data_t *afe);
/** /**
* @brief Initial wakenet and wake words coefficient, or reset wakenet and wake words coefficient * @brief Initial wakenet and wake words coefficient, or reset wakenet and wake words coefficient
@ -108,7 +109,7 @@ typedef afe_fetch_mode_t (*esp_afe_sr_iface_op_fetch_t)(esp_afe_sr_data_t *afe,
* @param wakenet_word The wakenet word, should be DEFAULT_WAKE_WORD or EXTRA_WAKE_WORD * @param wakenet_word The wakenet word, should be DEFAULT_WAKE_WORD or EXTRA_WAKE_WORD
* @return 0: fail, 1: success * @return 0: fail, 1: success
*/ */
typedef int (*esp_afe_sr_iface_op_set_wakenet_t)(esp_afe_sr_data_t *afe, esp_wn_word_t wakenet_word); typedef int (*esp_afe_sr_iface_op_set_wakenet_t)(esp_afe_sr_data_t *afe, char* model_name);
/** /**
* @brief Disable wakenet model. * @brief Disable wakenet model.
@ -170,7 +171,6 @@ typedef void (*esp_afe_sr_iface_op_destroy_t)(esp_afe_sr_data_t *afe);
* This structure contains the functions used to do operations on a AFE_SR. * This structure contains the functions used to do operations on a AFE_SR.
*/ */
typedef struct { typedef struct {
esp_afe_sr_iface_op_create_t create;
esp_afe_sr_iface_op_create_from_config_t create_from_config; esp_afe_sr_iface_op_create_from_config_t create_from_config;
esp_afe_sr_iface_op_feed_t feed; esp_afe_sr_iface_op_feed_t feed;
esp_afe_sr_iface_op_fetch_t fetch; esp_afe_sr_iface_op_fetch_t fetch;

19
include/esp32s3/esp_afe_sr_models.h
View File

@ -1,18 +1,27 @@
#pragma once #pragma once
#if defined CONFIG_USE_AFE #if defined CONFIG_USE_AFE
#include "esp_afe_sr_iface.h"
#if CONFIG_AFE_INTERFACE_V1 #if CONFIG_AFE_INTERFACE_V1
#include "esp_afe_sr_iface.h" extern const esp_afe_sr_iface_t esp_afe_sr_v1;
extern const esp_afe_sr_iface_t esp_afe_v1; extern const esp_afe_sr_iface_t esp_afe_voip_v1;
#define ESP_AFE_HANDLE esp_afe_v1 #define ESP_AFE_SR_HANDLE esp_afe_sr_v1
#define ESP_AFE_VOIP_HANDLE esp_afe_voip_v1
#else #else
#error No valid afe selected. #error No valid afe selected.
#endif #endif
#else #else
#include "esp_afe_sr_iface.h" #include "esp_afe_sr_iface.h"
extern const esp_afe_sr_iface_t esp_afe_v1; extern const esp_afe_sr_iface_t esp_afe_sr_v1;
#define ESP_AFE_HANDLE esp_afe_v1 extern const esp_afe_sr_iface_t esp_afe_voip_v1;
#define ESP_AFE_SR_HANDLE esp_afe_sr_v1
#define ESP_AFE_VOIP_HANDLE esp_afe_voip_v1
#endif #endif

96
include/esp32s3/esp_mn_iface.h
View File

@ -1,40 +1,53 @@
#pragma once #pragma once
#include "stdint.h" #include "stdint.h"
#include "dl_lib_coefgetter_if.h"
#include "esp_wn_iface.h" #include "esp_wn_iface.h"
// Return all possible recognition results
#define ESP_MN_RESULT_MAX_NUM 5 #define ESP_MN_RESULT_MAX_NUM 5
#define ESP_MN_MAX_PHRASE_NUM 200 #define ESP_MN_MAX_PHRASE_NUM 200
#define ESP_MN_MAX_PHRASE_LEN 63
#define ESP_MN_MIN_PHRASE_LEN 2
typedef enum { typedef enum {
ESP_MN_STATE_DETECTING = 0, // detecting ESP_MN_STATE_DETECTING = 0, // detecting
ESP_MN_STATE_DETECTED = 1, // detected ESP_MN_STATE_DETECTED = 1, // detected
ESP_MN_STATE_TIMEOUT = 2, // time out ESP_MN_STATE_TIMEOUT = 2, // time out
} esp_mn_state_t; } esp_mn_state_t;
// Return all possible recognition results
typedef struct{ typedef struct{
esp_mn_state_t state; esp_mn_state_t state;
int num; // The number of phrase in list, num<=5. When num=0, no phrase is recognized. int num; // The number of phrase in list, num<=5. When num=0, no phrase is recognized.
int command_id[ESP_MN_RESULT_MAX_NUM]; // The list of command id. int command_id[ESP_MN_RESULT_MAX_NUM]; // The list of command id.
int phrase_id[ESP_MN_RESULT_MAX_NUM]; // The list of phrase id. int phrase_id[ESP_MN_RESULT_MAX_NUM]; // The list of phrase id.
float prob[ESP_MN_RESULT_MAX_NUM]; // The list of probability. float prob[ESP_MN_RESULT_MAX_NUM]; // The list of probability.
} esp_mn_results_t; } esp_mn_results_t;
typedef struct{ typedef struct{
int16_t err_id_num; int16_t num; // The number of error phrases, which can not added into model
int16_t err_id[ESP_MN_MAX_PHRASE_NUM]; int16_t phrase_idx[ESP_MN_MAX_PHRASE_NUM]; // The error phrase index in singly linked list．
} esp_mn_phrase_err_id_t; } esp_mn_error_t;
typedef struct {
char phoneme_string[ESP_MN_MAX_PHRASE_LEN + 1]; // phoneme string
int16_t command_id; // the command id
float threshold; // trigger threshold, default: 0
int16_t *wave; // prompt wave data of the phrase
} esp_mn_phrase_t;
typedef struct _mn_node_ {
esp_mn_phrase_t *phrase;
struct _mn_node_ *next;
} esp_mn_node_t;
/** /**
* @brief Initialze a model instance with specified model coefficient. * @brief Initialze a model instance with specified model name.
* *
* @param coeff The wakenet model coefficient. * @param model_name The wakenet model name.
* @param duration The duration (ms) to trigger the timeout * @param duration The duration (ms) to trigger the timeout
* @parm sample_length Audio length for speech recognition, in ms. *
* @returns Handle to the model data. * @returns Handle to the model data.
*/ */
typedef model_iface_data_t* (*esp_mn_iface_op_create_t)(const model_coeff_getter_t *coeff, int duration); typedef model_iface_data_t* (*esp_mn_iface_op_create_t)(const char *model_name, int duration);
/** /**
* @brief Callback function type to fetch the amount of samples that need to be passed to the detect function * @brief Callback function type to fetch the amount of samples that need to be passed to the detect function
@ -62,24 +75,6 @@ typedef int (*esp_mn_iface_op_get_samp_chunknum_t)(model_iface_data_t *model);
* @param det_treshold The threshold to trigger speech commands, the range of det_threshold is 0.0~0.9999 * @param det_treshold The threshold to trigger speech commands, the range of det_threshold is 0.0~0.9999
*/ */
typedef int (*esp_mn_iface_op_set_det_threshold_t)(model_iface_data_t *model, float det_threshold); typedef int (*esp_mn_iface_op_set_det_threshold_t)(model_iface_data_t *model, float det_threshold);
/**
* @brief Set the detection threshold to manually abjust the probability
*
* @param model The model object to query
* @param phrase_id The ID of speech command phrase
* @param det_treshold The threshold to trigger speech command phrases
*/
typedef void (*esp_mn_iface_op_set_command_det_threshold_t)(model_iface_data_t *model, int phrase_id, float det_threshold);
/**
* @brief Get the detection threshold by phrase ID
*
* @param model The model object to query
* @param phrase_id The ID of speech command phrase
*
* @return The threshold of speech command phrases
*/
typedef float (*esp_mn_iface_op_get_command_det_threshold_t)(model_iface_data_t *model, int phrase_id);
/** /**
* @brief Get the sample rate of the samples to feed to the detect function * @brief Get the sample rate of the samples to feed to the detect function
@ -95,18 +90,9 @@ typedef int (*esp_mn_iface_op_get_samp_rate_t)(model_iface_data_t *model);
* @param model The model object to query. * @param model The model object to query.
* @param samples An array of 16-bit signed audio samples. The array size used can be queried by the * @param samples An array of 16-bit signed audio samples. The array size used can be queried by the
* get_samp_chunksize function. * get_samp_chunksize function.
* @return The command id, return 0 if no command word is detected, * @return The state of multinet
*/ */
typedef int (*esp_mn_iface_op_detect_t)(model_iface_data_t *model, int16_t *samples); typedef esp_mn_state_t (*esp_mn_iface_op_detect_t)(model_iface_data_t *model, int16_t *samples);
/**
* @brief Get the phrase if command word is detected
*
* @param model The model object to query.
*
* @return The phrase id, return the phrase if command word is detected.
*/
typedef int (*esp_mn_iface_op_get_det_phrase_id_t)(model_iface_data_t *model);
/** /**
* @brief Destroy a speech commands recognition model * @brief Destroy a speech commands recognition model
@ -115,17 +101,6 @@ typedef int (*esp_mn_iface_op_get_det_phrase_id_t)(model_iface_data_t *model);
*/ */
typedef void (*esp_mn_iface_op_destroy_t)(model_iface_data_t *model); typedef void (*esp_mn_iface_op_destroy_t)(model_iface_data_t *model);
/**
* @brief Reset the speech commands
*
* @param model_data The model object to query.
* @param command_str The string of new commands.
* @param command_id -1 means modify all commands with new command_str, 0-200 means modify the corresponding command
*
* @return The error ID structure
*/
typedef esp_mn_phrase_err_id_t* (*esp_mn_iface_op_reset_t)(model_iface_data_t *model_data, char *command_str, int command_id);
/** /**
* @brief Get recognition results * @brief Get recognition results
* *
@ -136,20 +111,21 @@ typedef esp_mn_phrase_err_id_t* (*esp_mn_iface_op_reset_t)(model_iface_data_t *m
typedef esp_mn_results_t* (*esp_mn_iface_op_get_results_t)(model_iface_data_t *model); typedef esp_mn_results_t* (*esp_mn_iface_op_get_results_t)(model_iface_data_t *model);
/** /**
* @brief Reset the speech commands recognition model * @brief Open the log print
* *
* @param model_data The model object to query. * @param model_data The model object to query.
* *
*/ */
typedef void (*esp_mn_iface_op_wakenet_reset_t)(model_iface_data_t *model_data); typedef void (*esp_mn_iface_op_open_log_t)(model_iface_data_t *model_data);
/** /**
* @brief Close the log print * @brief Set the speech commands by mn_command_root
* *
* @param model_data The model object to query. * @param model_data The model object to query.
* * @param mn_command_root The speech commands link.
* @return The error phrase id info.
*/ */
typedef void (*esp_mn_iface_op_close_log_t)(model_iface_data_t *model_data); typedef esp_mn_error_t* (*esp_wn_iface_op_set_speech_commands)(model_iface_data_t *model_data, esp_mn_node_t *mn_command_root);
typedef struct { typedef struct {
esp_mn_iface_op_create_t create; esp_mn_iface_op_create_t create;
@ -157,13 +133,9 @@ typedef struct {
esp_mn_iface_op_get_samp_chunksize_t get_samp_chunksize; esp_mn_iface_op_get_samp_chunksize_t get_samp_chunksize;
esp_mn_iface_op_get_samp_chunknum_t get_samp_chunknum; esp_mn_iface_op_get_samp_chunknum_t get_samp_chunknum;
esp_mn_iface_op_set_det_threshold_t set_det_threshold; esp_mn_iface_op_set_det_threshold_t set_det_threshold;
esp_mn_iface_op_set_command_det_threshold_t set_command_det_threshold;
esp_mn_iface_op_get_command_det_threshold_t get_command_det_threshold;
esp_mn_iface_op_get_det_phrase_id_t get_det_phrase_id;
esp_mn_iface_op_detect_t detect; esp_mn_iface_op_detect_t detect;
esp_mn_iface_op_destroy_t destroy; esp_mn_iface_op_destroy_t destroy;
esp_mn_iface_op_reset_t reset;
esp_mn_iface_op_get_results_t get_results; esp_mn_iface_op_get_results_t get_results;
esp_mn_iface_op_wakenet_reset_t wakenet_reset; esp_mn_iface_op_open_log_t open_log;
esp_mn_iface_op_close_log_t close_log; esp_wn_iface_op_set_speech_commands set_speech_commands;
} esp_mn_iface_t; } esp_mn_iface_t;

67
include/esp32s3/esp_mn_models.h
View File

@ -3,14 +3,28 @@
//Contains declarations of all available speech recognion models. Pair this up with the right coefficients and you have a model that can recognize //Contains declarations of all available speech recognion models. Pair this up with the right coefficients and you have a model that can recognize
//a specific phrase or word. //a specific phrase or word.
extern const esp_mn_iface_t esp_sr_multinet1_single_quantized_en;
extern const esp_mn_iface_t esp_sr_multinet3_single_quantized_en;
extern const esp_mn_iface_t esp_sr_multinet2_single_quantized_cn;
extern const esp_mn_iface_t esp_sr_multinet3_single_quantized_cn; #define ESP_MN_PREFIX "mn"
extern const esp_mn_iface_t esp_sr_multinet4_single_quantized_cn; #define ESP_MN_ENGLISH "en"
extern const esp_mn_iface_t esp_sr_multinet3_continuous_quantized_cn; #define ESP_MN_CHINESE "cn"
extern const esp_mn_iface_t esp_sr_multinet5_quantized;
extern const esp_mn_iface_t esp_sr_multinet5_quantized8; /**
* @brief Get the multinet handle from model name
*
* @param model_name The name of model
* @returns The handle of multinet
*/
esp_mn_iface_t *esp_mn_handle_from_name(char *model_name);
/**
* @brief Get the multinet language from model name
*
* @param model_name The name of model
* @returns The language of multinet
*/
char *esp_mn_language_from_name(char *model_name);
/* /*
Configure wake word to use based on what's selected in menuconfig. Configure wake word to use based on what's selected in menuconfig.
@ -19,42 +33,19 @@ extern const esp_mn_iface_t esp_sr_multinet5_quantized8;
#if CONFIG_SR_MN_CN_MULTINET2_SINGLE_RECOGNITION #if CONFIG_SR_MN_CN_MULTINET2_SINGLE_RECOGNITION
#include "multinet2_ch.h" #include "multinet2_ch.h"
#define MULTINET_MODEL_CN esp_sr_multinet2_single_quantized_cn
#define MULTINET_COEFF_CN get_coeff_multinet2_ch
#define MULTINET_MODEL esp_sr_multinet2_single_quantized_cn
#define MULTINET_COEFF get_coeff_multinet2_ch #define MULTINET_COEFF get_coeff_multinet2_ch
#elif CONFIG_SR_MN_CN_MULTINET4_SINGLE_RECOGNITION #define MULTINET_MODEL_NAME "mn2_cn"
#define MULTINET_MODEL_CN esp_sr_multinet4_single_quantized_cn
#define MULTINET_COEFF_CN "mn4cn"
#define MULTINET_MODEL esp_sr_multinet4_single_quantized_cn
#define MULTINET_COEFF "mn4cn"
#elif CONFIG_SR_MN_CN_MULTINET4_5_SINGLE_RECOGNITION
#define MULTINET_MODEL_CN esp_sr_multinet4_single_quantized_cn
#define MULTINET_COEFF_CN "mn4_5cn"
#define MULTINET_MODEL esp_sr_multinet4_single_quantized_cn
#define MULTINET_COEFF "mn4_5cn"
#endif
#if CONFIG_SR_MN_EN_MULTINET5_SINGLE_RECOGNITION #else
#define MULTINET_MODEL_EN esp_sr_multinet5_quantized #define MULTINET_COEFF "COEFF_NULL"
#define MULTINET_COEFF_EN "mn5en" #define MULTINET_MODEL_NAME "NULL"
#ifndef MULTINET_MODEL
#define MULTINET_MODEL esp_sr_multinet5_quantized
#define MULTINET_COEFF "mn5en"
#endif
#elif CONFIG_SR_MN_EN_MULTINET5_SINGLE_RECOGNITION_QUANT8
#define MULTINET_MODEL_EN esp_sr_multinet5_quantized8
#define MULTINET_COEFF_EN "mn5q8en"
#ifndef MULTINET_MODEL
#define MULTINET_MODEL esp_sr_multinet5_quantized8
#define MULTINET_COEFF "mn5q8en"
#endif
#endif #endif
#else #else
#define MULTINET_MODEL "NULL" #define MULTINET_COEFF "COEFF_NULL"
#define MULTINET_COEFF "NULL" #define MULTINET_MODEL_NAME "NULL"
#endif #endif
/* example /* example
static const esp_mn_iface_t *multinet = &MULTINET_MODEL; static const esp_mn_iface_t *multinet = &MULTINET_MODEL;

158
include/esp32s3/esp_mn_speech_commands.h Normal file
View File

@ -0,0 +1,158 @@
// Copyright 2015-2022 Espressif Systems (Shanghai) PTE LTD
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#pragma once
#include "esp_err.h"
#include "esp_mn_iface.h"
/*
esp_mn_node_t is a singly linked list which is used to manage speech commands.
It is easy to add one speech command into linked list and remove one speech command from linked list.
*/
/**
* @brief Initialze the speech commands singly linked list.
*
* @return
* - ESP_OK Success
* - ESP_ERR_NO_MEM No memory
* - ESP_ERR_INVALID_STATE The Speech Commands link has been initialized
*/
esp_err_t esp_mn_commands_alloc(void);
/**
* @brief Clear the speech commands linked list and free root node.
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE The Speech Commands link has not been initialized
*/
esp_err_t esp_mn_commands_free(void);
/**
* @brief Add one speech commands with phoneme string and command ID
*
* @param command_id The command ID
* @param phoneme_string The phoneme string of the speech commands
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE Fail
*/
esp_err_t esp_mn_commands_add(int command_id, char *phoneme_string);
/**
* @brief Modify one speech commands with new phoneme string
*
* @param old_phoneme_string The old phoneme string of the speech commands
* @param new_phoneme_string The new phoneme string of the speech commands
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE Fail
*/
esp_err_t esp_mn_commands_modify(char *old_phoneme_string, char *new_phoneme_string);
/**
* @brief Remove one speech commands by phoneme string
*
* @param phoneme_string The phoneme string of the speech commands
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE Fail
*/
esp_err_t esp_mn_commands_remove(char *phoneme_string);
/**
* @brief Clear all speech commands in linked list
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE Fail
*/
esp_err_t esp_mn_commands_clear(void);
/**
* @brief Get phrase from index, which is the depth from the phrase node to root node
*
* @Warning: The first phrase index is 0, the second phrase index is 1, and so on.
*
* @return
* - esp_mn_phrase_t* Success
* - NULL Fail
*/
esp_mn_phrase_t *esp_mn_commands_get_from_index(int index);
/**
* @brief Get phrase from phoneme string
*
* @return
* - esp_mn_phrase_t* Success
* - NULL Fail
*/
esp_mn_phrase_t *esp_mn_commands_get_from_string(const char *phoneme_string);
/**
* @brief Update the speech commands of MultiNet
*
* @Warning: Must be used after [add/remove/modify/clear] function,
* otherwise the language model of multinet can not be updated.
*
* @param multinet The multinet handle
* @param model_data The model object to query
*
* @return
* - NULL Success
* - others The list of error phrase which can not be parsed by multinet.
*/
esp_mn_error_t *esp_mn_commands_update(const esp_mn_iface_t *multinet, const model_iface_data_t *model_data);
/**
* @brief Print the MultiNet Speech Commands.
*/
void esp_mn_print_commands(void);
/**
* @brief Initialze the esp_mn_phrase_t struct by command id and phoneme string .
*
* @return the pointer of esp_mn_phrase_t
*/
esp_mn_phrase_t *esp_mn_phrase_alloc(int command_id, char *phoneme_string);
/**
* @brief Free esp_mn_phrase_t pointer.
*
* @param phrase The esp_mn_phrase_t pointer
*/
void esp_mn_phrase_free(esp_mn_phrase_t *phrase);
/**
* @brief Initialze the esp_mn_node_t struct by esp_mn_phrase_t pointer.
*
* @return the pointer of esp_mn_node_t
*/
esp_mn_node_t *esp_mn_node_alloc(esp_mn_phrase_t *phrase);
/**
* @brief Free esp_mn_node_t pointer.
*
* @param node The esp_mn_node_free pointer
*/
void esp_mn_node_free(esp_mn_node_t *node);
/**
* @brief Print phrase linked list.
*/
void esp_mn_commands_print(void);

23
include/esp32s3/esp_process_sdkconfig.h Normal file
View File

@ -0,0 +1,23 @@
#pragma once
#include "esp_err.h"
#include "esp_mn_iface.h"
/**
* @brief Check chip config to ensure optimum performance
*/
void check_chip_config(void);
/**
* @brief Update the speech commands of MultiNet by menuconfig
*
* @param multinet The multinet handle
*
* @param model_data The model object to query
*
* @param langugae The language of MultiNet
*
* @return
* - ESP_OK Success
* - ESP_ERR_INVALID_STATE Fail
*/
esp_err_t esp_mn_commands_update_from_sdkconfig(esp_mn_iface_t *multinet, const model_iface_data_t *model_data);

33
include/esp32s3/esp_wn_iface.h
View File

@ -1,10 +1,19 @@
#pragma once #pragma once
#include "stdint.h" #include "stdint.h"
#include "dl_lib_coefgetter_if.h"
//Opaque model data container //Opaque model data container
typedef struct model_iface_data_t model_iface_data_t; typedef struct model_iface_data_t model_iface_data_t;
/**
* @brief The state of wakeup
*/
typedef enum
{
WAKENET_NO_DETECT = 0, // wake word is not detected
WAKENET_CHANNEL_VERIFIED = -1, // output channel is verified
WAKENET_DETECTED = 1 // wake word is detected
} wakenet_state_t;
//Set wake words recognition operating mode //Set wake words recognition operating mode
//The probability of being wake words is increased with increasing mode, //The probability of being wake words is increased with increasing mode,
//As a consequence also the false alarm rate goes up //As a consequence also the false alarm rate goes up
@ -25,14 +34,14 @@ typedef struct {
/** /**
* @brief Easy function type to initialze a model instance with a detection mode and specified wake word coefficient * @brief Easy function type to initialze a model instance with a detection mode and specified wake word coefficient
* *
* @param model_name The specified wake word model coefficient
* @param det_mode The wake words detection mode to trigger wake words, DET_MODE_90 or DET_MODE_95 * @param det_mode The wake words detection mode to trigger wake words, DET_MODE_90 or DET_MODE_95
* @param model_coeff The specified wake word model coefficient
* @returns Handle to the model data * @returns Handle to the model data
*/ */
typedef model_iface_data_t* (*esp_wn_iface_op_create_t)(const model_coeff_getter_t *model_coeff, det_mode_t det_mode); typedef model_iface_data_t* (*esp_wn_iface_op_create_t)(const void *model_name, det_mode_t det_mode);
/** /**
* @brief Callback function type to fetch the amount of samples that need to be passed to the detect function * @brief Get the amount of samples that need to be passed to the detect function
* *
* Every speech recognition model processes a certain number of samples at the same time. This function * Every speech recognition model processes a certain number of samples at the same time. This function
* can be used to query that amount. Note that the returned amount is in 16-bit samples, not in bytes. * can be used to query that amount. Note that the returned amount is in 16-bit samples, not in bytes.
@ -43,7 +52,7 @@ typedef model_iface_data_t* (*esp_wn_iface_op_create_t)(const model_coeff_getter
typedef int (*esp_wn_iface_op_get_samp_chunksize_t)(model_iface_data_t *model); typedef int (*esp_wn_iface_op_get_samp_chunksize_t)(model_iface_data_t *model);
/** /**
* @brief Callback function type to fetch the channel number of samples that need to be passed to the detect function * @brief Get the channel number of samples that need to be passed to the detect function
* *
* Every speech recognition model processes a certain number of samples at the same time. This function * Every speech recognition model processes a certain number of samples at the same time. This function
* can be used to query that amount. Note that the returned amount is in 16-bit samples, not in bytes. * can be used to query that amount. Note that the returned amount is in 16-bit samples, not in bytes.
@ -53,6 +62,17 @@ typedef int (*esp_wn_iface_op_get_samp_chunksize_t)(model_iface_data_t *model);
*/ */
typedef int (*esp_wn_iface_op_get_channel_num_t)(model_iface_data_t *model); typedef int (*esp_wn_iface_op_get_channel_num_t)(model_iface_data_t *model);
/**
* @brief Get the start point of wake word when one wake word is detected.
*
* @Warning: This function should be called when the channel index is verified.
* The returned value is the number of samples from start point of wake word to detected point.
*
* @param model The model object to query
* @return The number of samples from start point to detected point (end point)
*/
typedef int (*esp_wn_iface_op_get_start_point_t)(model_iface_data_t *model);
/** /**
* @brief Get the sample rate of the samples to feed to the detect function * @brief Get the sample rate of the samples to feed to the detect function
@ -110,7 +130,7 @@ typedef float (*esp_wn_iface_op_get_det_threshold_t)(model_iface_data_t *model,
* get_samp_chunksize function. * get_samp_chunksize function.
* @return The index of wake words, return 0 if no wake word is detected, else the index of the wake words. * @return The index of wake words, return 0 if no wake word is detected, else the index of the wake words.
*/ */
typedef int (*esp_wn_iface_op_detect_t)(model_iface_data_t *model, int16_t *samples); typedef wakenet_state_t (*esp_wn_iface_op_detect_t)(model_iface_data_t *model, int16_t *samples);
/** /**
* @brief Get the volume gain * @brief Get the volume gain
@ -149,6 +169,7 @@ typedef void (*esp_wn_iface_op_destroy_t)(model_iface_data_t *model);
*/ */
typedef struct { typedef struct {
esp_wn_iface_op_create_t create; esp_wn_iface_op_create_t create;
esp_wn_iface_op_get_start_point_t get_start_point;
esp_wn_iface_op_get_samp_chunksize_t get_samp_chunksize; esp_wn_iface_op_get_samp_chunksize_t get_samp_chunksize;
esp_wn_iface_op_get_channel_num_t get_channel_num; esp_wn_iface_op_get_channel_num_t get_channel_num;
esp_wn_iface_op_get_samp_rate_t get_samp_rate; esp_wn_iface_op_get_samp_rate_t get_samp_rate;

162
include/esp32s3/esp_wn_models.h
View File

@ -1,141 +1,109 @@
#pragma once #pragma once
#include "esp_wn_iface.h" #include "esp_wn_iface.h"
//Contains declarations of all available speech recognion models. Pair this up with the right coefficients and you have a model that can recognize
//a specific phrase or word.
typedef enum { // The prefix of wakenet model name is used to filter all wakenet from availabel models.
FIRST_WAKE_WORD = 0, #define ESP_WN_PREFIX "wn"
SECOND_WAKE_WORD,
} esp_wn_word_t;
typedef enum { /**
NO_WAKE_WORD = -1, * @brief Get the wakenet handle from model name
CUSTOMIZED_WORD_WN7 = 0, *
XIAOAITONGXUE_WN7 = 1, * @param model_name The name of model
XIAOAITONGXUE_WN7Q8, * @returns The handle of wakenet
ALEXA_WN7, */
ALEXA_WN7Q8, const esp_wn_iface_t *esp_wn_handle_from_name(const char *model_name);
ALEXA_WN8,
HILEXIN_WN7,
HILEXIN_WN7Q8,
HILEXIN_WN8,
HIESP_WN8,
HIESP_WN8Q8,
} esp_wn_model_t;
extern const esp_wn_iface_t esp_sr_wakenet5_quantized; /**
extern const esp_wn_iface_t esp_sr_wakenet7_quantized; * @brief Get the wake word name from model name
extern const esp_wn_iface_t esp_sr_wakenet7_quantized8; *
extern const esp_wn_iface_t esp_sr_wakenet8_quantized; * @param model_name The name of model
extern const esp_wn_iface_t esp_sr_wakenet8_quantized8; * @returns The wake word name, like "alexa","hilexin","xiaoaitongxue"
*/
char* esp_wn_wakeword_from_name(const char *model_name);
// /**
// * @brief Get the model coeff from model name
// *
// * @Warning: retuen model_coeff_getter_t, when chip is ESP32,
// * return string for other chips
// *
// * @param model_name The name of model
// * @returns The handle of wakenet
// */
// void *esp_wn_coeff_from_name(char *model_name);
#if defined CONFIG_USE_WAKENET
/* /*
Configure wake word to use based on what's selected in menuconfig. Configure wake word to use based on what's selected in menuconfig.
*/ */
#if defined CONFIG_USE_WAKENET #if CONFIG_SR_WN_WN5_HILEXIN
#if CONFIG_SR_WN_WN5_HILEXIN_FIRST || CONFIG_SR_WN_WN5_HILEXIN_SECOND
#include "hilexin_wn5.h" #include "hilexin_wn5.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_hilexin"
#define WAKENET_COEFF get_coeff_hilexin_wn5 #define WAKENET_COEFF get_coeff_hilexin_wn5
#elif CONFIG_SR_WN_WN5X2_HILEXIN_FIRST || CONFIG_SR_WN_WN5X2_HILEXIN_SECOND #elif CONFIG_SR_WN_WN5X2_HILEXIN
#include "hilexin_wn5X2.h" #include "hilexin_wn5X2.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_hilexinX2"
#define WAKENET_COEFF get_coeff_hilexin_wn5X2 #define WAKENET_COEFF get_coeff_hilexin_wn5X2
#elif CONFIG_SR_WN_WN5X3_HILEXIN_FIRST || CONFIG_SR_WN_WN5X3_HILEXIN_SECOND
#elif CONFIG_SR_WN_WN5X3_HILEXIN
#include "hilexin_wn5X3.h" #include "hilexin_wn5X3.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_hilexinX3"
#define WAKENET_COEFF get_coeff_hilexin_wn5X3 #define WAKENET_COEFF get_coeff_hilexin_wn5X3
#elif CONFIG_SR_WN_WN5_NIHAOXIAOZHI_FIRST || CONFIG_SR_WN_WN5_NIHAOXIAOZHI_SECOND
#elif CONFIG_SR_WN_WN5_NIHAOXIAOZHI
#include "nihaoxiaozhi_wn5.h" #include "nihaoxiaozhi_wn5.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_nihaoxiaozhi"
#define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5 #define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5
#elif CONFIG_SR_WN_WN5X2_NIHAOXIAOZHI_FIRST || CONFIG_SR_WN_WN5X2_NIHAOXIAOZHI_SECOND
#elif CONFIG_SR_WN_WN5X2_NIHAOXIAOZHI
#include "nihaoxiaozhi_wn5X2.h" #include "nihaoxiaozhi_wn5X2.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_nihaoxiaozhiX2"
#define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5X2 #define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5X2
#elif CONFIG_SR_WN_WN5X3_NIHAOXIAOZHI_FIRST || CONFIG_SR_WN_WN5X3_NIHAOXIAOZHI_SECOND
#elif CONFIG_SR_WN_WN5X3_NIHAOXIAOZHI
#include "nihaoxiaozhi_wn5X3.h" #include "nihaoxiaozhi_wn5X3.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_nihaoxiaozhiX3"
#define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5X3 #define WAKENET_COEFF get_coeff_nihaoxiaozhi_wn5X3
#elif CONFIG_SR_WN_WN5X3_NIHAOXIAOXIN_FIRST || CONFIG_SR_WN_WN5X3_NIHAOXIAOXIN_SECOND
#elif CONFIG_SR_WN_WN5X3_NIHAOXIAOXIN
#include "nihaoxiaoxin_wn5X3.h" #include "nihaoxiaoxin_wn5X3.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_nihaoxiaoxinX3"
#define WAKENET_COEFF get_coeff_nihaoxiaoxin_wn5X3 #define WAKENET_COEFF get_coeff_nihaoxiaoxin_wn5X3
#elif CONFIG_SR_WN_WN5X3_HIJESON_FIRST || CONFIG_SR_WN_WN5X3_HIJESON_SECOND
#elif CONFIG_SR_WN_WN5X3_HIJESON
#include "hijeson_wn5X3.h" #include "hijeson_wn5X3.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_hijesonX3"
#define WAKENET_COEFF get_coeff_hijeson_wn5X3 #define WAKENET_COEFF get_coeff_hijeson_wn5X3
#elif CONFIG_SR_WN_WN5_CUSTOMIZED_WORD_FIRST || CONFIG_SR_WN_WN5_CUSTOMIZED_WORD_SECOND #elif CONFIG_SR_WN_WN5_CUSTOMIZED_WORD
#include "customized_word_wn5.h" #include "customized_word_wn5.h"
#define WAKENET_MODEL esp_sr_wakenet5_quantized #define WAKENET_MODEL_NAME "wn5_customizedword"
#define WAKENET_COEFF get_coeff_customized_word_wn5 #define WAKENET_COEFF get_coeff_customizedword_wn5
#elif CONFIG_SR_WN_WN7_CUSTOMIZED_WORD_FIRST || CONFIG_SR_WN_WN7_CUSTOMIZED_WORD_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized
#define WAKENET_COEFF "custom7"
#elif CONFIG_SR_WN_WN7_XIAOAITONGXUE_FIRST || CONFIG_SR_WN_WN7_XIAOAITONGXUE_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized
#define WAKENET_COEFF "xiaoaitongxue7"
#elif CONFIG_SR_WN_WN7Q8_XIAOAITONGXUE_FIRST || CONFIG_SR_WN_WN7Q8_XIAOAITONGXUE_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized8
#define WAKENET_COEFF "xiaoaitongxue7q8"
#elif CONFIG_SR_WN_WN7_HILEXIN_FIRST || CONFIG_SR_WN_WN7_HILEXIN_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized
#define WAKENET_COEFF "hilexin7"
#elif CONFIG_SR_WN_WN7Q8_HILEXIN_FIRST || CONFIG_SR_WN_WN7Q8_HILEXIN_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized8
#define WAKENET_COEFF "hilexin7q8"
#elif CONFIG_SR_WN_WN7_ALEXA_FIRST || CONFIG_SR_WN_WN7_ALEXA_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized
#define WAKENET_COEFF "alexa7"
#elif CONFIG_SR_WN_WN8_ALEXA_FIRST || CONFIG_SR_WN_WN8_ALEXA_SECOND
#define WAKENET_MODEL esp_sr_wakenet8_quantized
#define WAKENET_COEFF "alexa8"
#elif CONFIG_SR_WN_WN7Q8_ALEXA_FIRST || CONFIG_SR_WN_WN7Q8_ALEXA_SECOND
#define WAKENET_MODEL esp_sr_wakenet7_quantized8
#define WAKENET_COEFF "alexa7q8"
#elif CONFIG_SR_WN_WN8_HIESP_FIRST || CONFIG_SR_WN_WN8_HIESP_SECOND
#define WAKENET_MODEL esp_sr_wakenet8_quantized
#define WAKENET_COEFF "hiesp8"
#elif CONFIG_SR_WN_WN8Q8_HIESP_FIRST || CONFIG_SR_WN_WN8Q8_HIESP_SECOND
#define WAKENET_MODEL esp_sr_wakenet8_quantized8
#define WAKENET_COEFF "hiesp8q8"
#elif CONFIG_SR_WN_WN8_HILEXIN_FIRST || CONFIG_SR_WN_WN8_HILEXIN_SECOND
#define WAKENET_MODEL esp_sr_wakenet8_quantized
#define WAKENET_COEFF "hilexin8"
#else #else
#error No valid wake word selected. #define WAKENET_MODEL_NAME "NULL"
#define WAKENET_COEFF "COEFF_NULL"
#endif #endif
#else
#define WAKENET_MODEL "NULL"
#define WAKENET_COEFF "NULL"
#endif
/* example
static const sr_model_iface_t *model = &WAKENET_MODEL; #else
#define WAKENET_MODEL_NAME "NULL"
#define WAKENET_COEFF "COEFF_NULL"
#endif
/*
static const sr_model_iface_t *model = esp_wn_handle_from_name(model_name);
//Initialize wakeNet model data //Initialize wakeNet model data
static model_iface_data_t *model_data=model->create(DET_MODE_90); static model_iface_data_t *model_data=model->create(model_name, DET_MODE_90);
//Set parameters of buffer //Set parameters of buffer
int audio_chunksize=model->get_samp_chunksize(model_data); int audio_chunksize=model->get_samp_chunksize(model_data);

75
include/esp32s3/model_path.h
View File

@ -1,3 +1,74 @@
#pragma once #pragma once
char *get_model_base_path(void);
void srmodel_spiffs_init(void); typedef struct
{
char **model_name; // the name of models, like "wn9_hilexin"(wakenet9, hilexin), "mn5_en"(multinet5, english)
int num; // the number of models
} srmodel_list_t;
#define MODEL_NAME_MAX_LENGTH 64
/**
* @brief Return all avaliable models in spiffs or selected in Kconfig.
*
* @return all avaliable models in spiffs,save as srmodel_list_t.
*/
srmodel_list_t* esp_srmodel_init(void);
/**
* @brief Free srmodel_list_t and unregister SPIFFS filesystem if open SPIFFS filesystem.
*
* @param models The srmodel_list_t point allocated by esp_srmodel_init function.
*
* @return all avaliable models in spiffs,save as srmodel_list_t.
*/
void esp_srmodel_deinit(srmodel_list_t *models);
/**
* @brief Return the first model name containing the specified keywords
* If keyword is NULL, we will ignore the keyword.
*
* @param models The srmodel_list_t point allocated by esp_srmodel_init function.
* @param keyword1 The specified keyword1 , like ESP_WN_PREDIX(the prefix of wakenet),
* ESP_MN_PREFIX(the prefix of multinet),
*
* @param keyword2 The specified keyword2, like ESP_MN_ENGLISH(the english multinet)
* ESP_MN_CHINESE(the chinese multinet)
* "alexa" (the "alexa" wakenet)
* @return return model name if can find one model name containing the keywords otherwise return NULL.
*/
char *esp_srmodel_filter(srmodel_list_t *models, const char *keyword1, const char *keyword2);
/**
* @brief Check whether the specified model name exists or not.
*
* @param models The srmodel_list_t point allocated by esp_srmodel_init function.
* @param model_name The specified model name
* @return return index in models if model name exists otherwise return -1
*/
int esp_srmodel_exists(srmodel_list_t *models, char *model_name);
/**
* @brief Initialize and mount SPIFFS filesystem, return all avaliable models in spiffs.
*
* @return all avaliable models in spiffs,save as srmodel_list_t.
*/
srmodel_list_t *srmodel_spiffs_init(void);
/**
* @brief unregister SPIFFS filesystem and free srmodel_list_t.
*
* @param models The srmodel_list_t point allocated by srmodel_spiffs_init function.
*
* @return all avaliable models in spiffs,save as srmodel_list_t.
*/
void srmodel_spiffs_deinit(srmodel_list_t *models);
/**
* @brief Return base path of srmodel spiffs
*
* @return the base path od srmodel spiffs
*/
char *get_model_base_path(void);

BIN
lib/esp32s3/libc_speech_features.a
View File

Binary file not shown.

BIN
lib/esp32s3/libdl_lib.a
View File

Binary file not shown.

BIN
lib/esp32s3/libesp_audio_front_end.a
View File

Binary file not shown.

BIN
lib/esp32s3/libesp_audio_processor.a
View File

Binary file not shown.

BIN
lib/esp32s3/libmultinet.a
View File

Binary file not shown.

BIN
lib/esp32s3/libwakenet.a
View File

Binary file not shown.

242
src/esp_mn_speech_commands.c Normal file
View File

@ -0,0 +1,242 @@
#include <stdio.h>
#include <stdlib.h>
#include "string.h"
#include "esp_log.h"
#include "esp_heap_caps.h"
#include "esp_mn_speech_commands.h"
static char *TAG = "MN_COMMAND";
static esp_mn_node_t *esp_mn_root = NULL;
#define ESP_RETURN_ON_FALSE(a, err_code, log_tag, format, ...) do { \
if (!(a)) { \
ESP_LOGE(log_tag, "%s(%d): " format, __FUNCTION__, __LINE__, ##__VA_ARGS__); \
return err_code; \
} \
} while(0)
esp_err_t esp_mn_commands_alloc(void)
{
ESP_RETURN_ON_FALSE(NULL == esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands already initialized");
esp_mn_root = esp_mn_node_alloc(NULL);
return ESP_OK;
}
esp_err_t esp_mn_commands_free(void)
{
esp_mn_commands_clear();
esp_mn_node_free(esp_mn_root);
esp_mn_root = NULL;
return ESP_OK;
}
int esp_mn_commands_num(void)
{
esp_mn_node_t * t = esp_mn_root;
int i = 0;
while (t->next) {
t = t->next;
i++;
}
return i;
}
esp_err_t esp_mn_commands_clear(void)
{
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
esp_mn_node_t * t = esp_mn_root->next;
while (t) {
esp_mn_node_t *cur_node = t;
t = t->next;
esp_mn_node_free(cur_node);
}
esp_mn_root->next = NULL;
return ESP_OK;
}
esp_err_t esp_mn_commands_add(int command_id, char *phoneme_string)
{
esp_mn_node_t *temp = esp_mn_root;
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
int last_node_elem_num = esp_mn_commands_num();
ESP_RETURN_ON_FALSE(ESP_MN_MAX_PHRASE_NUM >= last_node_elem_num, ESP_ERR_INVALID_STATE, TAG, "The number of speech commands phrase must less than 200");
esp_mn_phrase_t *phrase = esp_mn_phrase_alloc(command_id, phoneme_string);
if (phrase == NULL) return ESP_ERR_INVALID_STATE;
esp_mn_node_t *new_node = esp_mn_node_alloc(phrase);
while (temp->next != NULL) {
temp = temp->next;
}
temp->next = new_node;
return ESP_OK;
}
esp_err_t esp_mn_commands_modify(char *old_phoneme_string, char *new_phoneme_string)
{
esp_mn_node_t *temp = esp_mn_root;
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
// search old phoneme_string to get command id
bool flag = false;
int command_id;
while (temp->next) {
temp = temp->next;
if (strcmp(old_phoneme_string, temp->phrase->phoneme_string) == 0) {
command_id = temp->phrase->command_id;
flag = true;
break;
}
}
// replace old phrase with new phrase
if (flag) {
esp_mn_phrase_t *phrase = esp_mn_phrase_alloc(command_id, new_phoneme_string);
if (phrase == NULL) return ESP_ERR_INVALID_STATE;
esp_mn_phrase_free(temp->phrase);
temp->phrase = phrase;
} else {
ESP_LOGE(TAG, "No such speech command: \"%s\"", old_phoneme_string);
return ESP_ERR_INVALID_STATE;
}
return ESP_OK;
}
esp_err_t esp_mn_commands_remove(char *phoneme_string)
{
esp_mn_node_t *temp = esp_mn_root;
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
// search phoneme_string to get node point
bool flag = false;
while (temp->next) {
if (strcmp(phoneme_string, temp->next->phrase->phoneme_string) == 0) {
flag = true;
break;
}
temp = temp->next;
}
if (flag) {
esp_mn_node_t *rm_node = temp->next;
temp->next = temp->next->next;
esp_mn_node_free(rm_node);
return ESP_OK;
} else {
ESP_LOGE(TAG, "No such speech command: \"%s\"", phoneme_string);
return ESP_ERR_INVALID_STATE;
}
return ESP_OK;
}
esp_mn_phrase_t *esp_mn_commands_get_from_index(int index)
{
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
// phrase index also is phrase id, which is the depth from this phrase node to root node
esp_mn_node_t *temp = esp_mn_root;
for (int i=-1; i<index; i++) {
if (temp->next == NULL) return NULL;
temp = temp->next;
}
return temp->phrase;
}
esp_mn_phrase_t *esp_mn_commands_get_from_string(const char *phoneme_string)
{
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialized");
// phrase index also is phrase id, which is the depth from this phrase node to root node
esp_mn_node_t *temp = esp_mn_root;
while (temp->next) {
if (strcmp(phoneme_string, temp->next->phrase->phoneme_string) == 0) {
return temp->next->phrase;
}
temp = temp->next;
}
return NULL;
}
esp_mn_error_t *esp_mn_commands_update(const esp_mn_iface_t *multinet, const model_iface_data_t *model_data)
{
ESP_RETURN_ON_FALSE(NULL != esp_mn_root, ESP_ERR_INVALID_STATE, TAG, "The mn commands is not initialize");
esp_mn_error_t *error = multinet->set_speech_commands(model_data, esp_mn_root);
if (error->num == 0)
return NULL;
return error;
}
void esp_mn_commands_print(void)
{
ESP_LOGI(TAG, "---------------------SPEECH COMMANDS---------------------");
esp_mn_node_t* temp = esp_mn_root;
int phrase_id = 0;
while (temp->next) {
temp = temp->next;
ESP_LOGI(TAG, "Command ID%d, phrase ID%d: %s", temp->phrase->command_id, phrase_id, temp->phrase->phoneme_string);
phrase_id++;
}
ESP_LOGI(TAG, "---------------------------------------------------------\n");
}
void *_esp_mn_calloc_(int n, int size)
{
#ifdef ESP_PLATFORM
return heap_caps_calloc(n, size, MALLOC_CAP_SPIRAM);
#else
return calloc(n, size);
#endif
}
esp_mn_phrase_t *esp_mn_phrase_alloc(int command_id, char *phoneme_string)
{
int phoneme_string_len = strlen(phoneme_string);
if (phoneme_string_len > ESP_MN_MAX_PHRASE_LEN || phoneme_string_len<1) {
ESP_LOGE(TAG, "The Length of \"%s\" > ESP_MN_MAX_PHRASE_LEN", phoneme_string);
return NULL;
}
esp_mn_phrase_t *phrase = _esp_mn_calloc_(1, sizeof(esp_mn_phrase_t));
ESP_RETURN_ON_FALSE(NULL != phrase, ESP_ERR_NO_MEM, TAG, "Fail to alloc mn phrase");
phrase->command_id = command_id;
phrase->threshold = 0;
phrase->wave = NULL;
memcpy(phrase->phoneme_string, phoneme_string, phoneme_string_len);
return phrase;
}
void esp_mn_phrase_free(esp_mn_phrase_t *phrase)
{
if (phrase != NULL) {
free(phrase);
}
}
esp_mn_node_t *esp_mn_node_alloc(esp_mn_phrase_t *phrase)
{
esp_mn_node_t *node = _esp_mn_calloc_(1, sizeof(esp_mn_node_t));
ESP_RETURN_ON_FALSE(NULL != node, ESP_ERR_NO_MEM, TAG, "Fail to alloc mn node");
node->next = NULL;
node->phrase = phrase;
return node;
}
void esp_mn_node_free(esp_mn_node_t *node)
{
if (node != NULL) {
esp_mn_phrase_free(node->phrase);
free(node);
}
}

924
src/esp_process_sdkconfig.c Normal file
View File

@ -0,0 +1,924 @@
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include "sdkconfig.h"
#include "esp_log.h"
#include "esp_process_sdkconfig.h"
#include "esp_mn_speech_commands.h"
static char *TAG = "SR_SYS";
void check_chip_config(void)
{
#ifdef CONFIG_IDF_TARGET_ESP32S3
#ifndef CONFIG_ESP32S3_DEFAULT_CPU_FREQ_240
ESP_LOGE(TAG, "CPU freq should be 240MHz");
#endif
#ifndef CONFIG_ESPTOOLPY_FLASHFREQ_80M
ESP_LOGE(TAG, "Flash freq should be 240MHz");
#endif
#ifndef CONFIG_SPIRAM_SPEED_80M
ESP_LOGE(TAG, "PSRAM freq should be 80MHz");
#endif
#ifndef CONFIG_ESP32S3_DATA_CACHE_64KB
ESP_LOGE(TAG, "Data cache should be 64KB");
#endif
#ifndef CONFIG_ESP32S3_DATA_CACHE_LINE_64B
ESP_LOGE(TAG, "Data cache line should be 64B");
#endif
#elif CONFIG_IDF_TARGET_ESP32
#ifndef CONFIG_ESP32_DEFAULT_CPU_FREQ_240
ESP_LOGE(TAG, "CPU freq should be 240MHz");
#endif
#ifndef CONFIG_SPIRAM_SPEED_80M
ESP_LOGE(TAG, "PSRAM freq should be 80MHz");
#endif
#ifndef CONFIG_ESPTOOLPY_FLASHFREQ_80M
ESP_LOGE(TAG, "Flash freq should be 80MHz");
#endif
#ifndef CONFIG_ESPTOOLPY_FLASHMODE_QIO
ESP_LOGE(TAG, "Flash mode should be QIO");
#endif
#else
ESP_LOGE(TAG, "ESP-SR-AFE only support ESP32/ESP32S3");
#endif
}
char *get_id_name_cn(int i)
{
#if defined CONFIG_USE_MULTINET
#if defined CONFIG_SR_MN_CN_MULTINET2_SINGLE_RECOGNITION || defined CONFIG_SR_MN_CN_MULTINET4_5_SINGLE_RECOGNITION || defined CONFIG_SR_MN_CN_MULTINET4_5_SINGLE_RECOGNITION_QUANT8
if (i == 0)
return CONFIG_CN_SPEECH_COMMAND_ID0;
else if (i == 1)
return CONFIG_CN_SPEECH_COMMAND_ID1;
else if (i == 2)
return CONFIG_CN_SPEECH_COMMAND_ID2;
else if (i == 3)
return CONFIG_CN_SPEECH_COMMAND_ID3;
else if (i == 4)
return CONFIG_CN_SPEECH_COMMAND_ID4;
else if (i == 5)
return CONFIG_CN_SPEECH_COMMAND_ID5;
else if (i == 6)
return CONFIG_CN_SPEECH_COMMAND_ID6;
else if (i == 7)
return CONFIG_CN_SPEECH_COMMAND_ID7;
else if (i == 8)
return CONFIG_CN_SPEECH_COMMAND_ID8;
else if (i == 9)
return CONFIG_CN_SPEECH_COMMAND_ID9;
else if (i == 10)
return CONFIG_CN_SPEECH_COMMAND_ID10;
else if (i == 11)
return CONFIG_CN_SPEECH_COMMAND_ID11;
else if (i == 12)
return CONFIG_CN_SPEECH_COMMAND_ID12;
else if (i == 13)
return CONFIG_CN_SPEECH_COMMAND_ID13;
else if (i == 14)
return CONFIG_CN_SPEECH_COMMAND_ID14;
else if (i == 15)
return CONFIG_CN_SPEECH_COMMAND_ID15;
else if (i == 16)
return CONFIG_CN_SPEECH_COMMAND_ID16;
else if (i == 17)
return CONFIG_CN_SPEECH_COMMAND_ID17;
else if (i == 18)
return CONFIG_CN_SPEECH_COMMAND_ID18;
else if (i == 19)
return CONFIG_CN_SPEECH_COMMAND_ID19;
else if (i == 20)
return CONFIG_CN_SPEECH_COMMAND_ID20;
else if (i == 21)
return CONFIG_CN_SPEECH_COMMAND_ID21;
else if (i == 22)
return CONFIG_CN_SPEECH_COMMAND_ID22;
else if (i == 23)
return CONFIG_CN_SPEECH_COMMAND_ID23;
else if (i == 24)
return CONFIG_CN_SPEECH_COMMAND_ID24;
else if (i == 25)
return CONFIG_CN_SPEECH_COMMAND_ID25;
else if (i == 26)
return CONFIG_CN_SPEECH_COMMAND_ID26;
else if (i == 27)
return CONFIG_CN_SPEECH_COMMAND_ID27;
else if (i == 28)
return CONFIG_CN_SPEECH_COMMAND_ID28;
else if (i == 29)
return CONFIG_CN_SPEECH_COMMAND_ID29;
else if (i == 30)
return CONFIG_CN_SPEECH_COMMAND_ID30;
else if (i == 31)
return CONFIG_CN_SPEECH_COMMAND_ID31;
else if (i == 32)
return CONFIG_CN_SPEECH_COMMAND_ID32;
else if (i == 33)
return CONFIG_CN_SPEECH_COMMAND_ID33;
else if (i == 34)
return CONFIG_CN_SPEECH_COMMAND_ID34;
else if (i == 35)
return CONFIG_CN_SPEECH_COMMAND_ID35;
else if (i == 36)
return CONFIG_CN_SPEECH_COMMAND_ID36;
else if (i == 37)
return CONFIG_CN_SPEECH_COMMAND_ID37;
else if (i == 38)
return CONFIG_CN_SPEECH_COMMAND_ID38;
else if (i == 39)
return CONFIG_CN_SPEECH_COMMAND_ID39;
else if (i == 40)
return CONFIG_CN_SPEECH_COMMAND_ID40;
else if (i == 41)
return CONFIG_CN_SPEECH_COMMAND_ID41;
else if (i == 42)
return CONFIG_CN_SPEECH_COMMAND_ID42;
else if (i == 43)
return CONFIG_CN_SPEECH_COMMAND_ID43;
else if (i == 44)
return CONFIG_CN_SPEECH_COMMAND_ID44;
else if (i == 45)
return CONFIG_CN_SPEECH_COMMAND_ID45;
else if (i == 46)
return CONFIG_CN_SPEECH_COMMAND_ID46;
else if (i == 47)
return CONFIG_CN_SPEECH_COMMAND_ID47;
else if (i == 48)
return CONFIG_CN_SPEECH_COMMAND_ID48;
else if (i == 49)
return CONFIG_CN_SPEECH_COMMAND_ID49;
else if (i == 50)
return CONFIG_CN_SPEECH_COMMAND_ID50;
else if (i == 51)
return CONFIG_CN_SPEECH_COMMAND_ID51;
else if (i == 52)
return CONFIG_CN_SPEECH_COMMAND_ID52;
else if (i == 53)
return CONFIG_CN_SPEECH_COMMAND_ID53;
else if (i == 54)
return CONFIG_CN_SPEECH_COMMAND_ID54;
else if (i == 55)
return CONFIG_CN_SPEECH_COMMAND_ID55;
else if (i == 56)
return CONFIG_CN_SPEECH_COMMAND_ID56;
else if (i == 57)
return CONFIG_CN_SPEECH_COMMAND_ID57;
else if (i == 58)
return CONFIG_CN_SPEECH_COMMAND_ID58;
else if (i == 59)
return CONFIG_CN_SPEECH_COMMAND_ID59;
else if (i == 60)
return CONFIG_CN_SPEECH_COMMAND_ID60;
else if (i == 61)
return CONFIG_CN_SPEECH_COMMAND_ID61;
else if (i == 62)
return CONFIG_CN_SPEECH_COMMAND_ID62;
else if (i == 63)
return CONFIG_CN_SPEECH_COMMAND_ID63;
else if (i == 64)
return CONFIG_CN_SPEECH_COMMAND_ID64;
else if (i == 65)
return CONFIG_CN_SPEECH_COMMAND_ID65;
else if (i == 66)
return CONFIG_CN_SPEECH_COMMAND_ID66;
else if (i == 67)
return CONFIG_CN_SPEECH_COMMAND_ID67;
else if (i == 68)
return CONFIG_CN_SPEECH_COMMAND_ID68;
else if (i == 69)
return CONFIG_CN_SPEECH_COMMAND_ID69;
else if (i == 70)
return CONFIG_CN_SPEECH_COMMAND_ID70;
else if (i == 71)
return CONFIG_CN_SPEECH_COMMAND_ID71;
else if (i == 72)
return CONFIG_CN_SPEECH_COMMAND_ID72;
else if (i == 73)
return CONFIG_CN_SPEECH_COMMAND_ID73;
else if (i == 74)
return CONFIG_CN_SPEECH_COMMAND_ID74;
else if (i == 75)
return CONFIG_CN_SPEECH_COMMAND_ID75;
else if (i == 76)
return CONFIG_CN_SPEECH_COMMAND_ID76;
else if (i == 77)
return CONFIG_CN_SPEECH_COMMAND_ID77;
else if (i == 78)
return CONFIG_CN_SPEECH_COMMAND_ID78;
else if (i == 79)
return CONFIG_CN_SPEECH_COMMAND_ID79;
else if (i == 80)
return CONFIG_CN_SPEECH_COMMAND_ID80;
else if (i == 81)
return CONFIG_CN_SPEECH_COMMAND_ID81;
else if (i == 82)
return CONFIG_CN_SPEECH_COMMAND_ID82;
else if (i == 83)
return CONFIG_CN_SPEECH_COMMAND_ID83;
else if (i == 84)
return CONFIG_CN_SPEECH_COMMAND_ID84;
else if (i == 85)
return CONFIG_CN_SPEECH_COMMAND_ID85;
else if (i == 86)
return CONFIG_CN_SPEECH_COMMAND_ID86;
else if (i == 87)
return CONFIG_CN_SPEECH_COMMAND_ID87;
else if (i == 88)
return CONFIG_CN_SPEECH_COMMAND_ID88;
else if (i == 89)
return CONFIG_CN_SPEECH_COMMAND_ID89;
else if (i == 90)
return CONFIG_CN_SPEECH_COMMAND_ID90;
else if (i == 91)
return CONFIG_CN_SPEECH_COMMAND_ID91;
else if (i == 92)
return CONFIG_CN_SPEECH_COMMAND_ID92;
else if (i == 93)
return CONFIG_CN_SPEECH_COMMAND_ID93;
else if (i == 94)
return CONFIG_CN_SPEECH_COMMAND_ID94;
else if (i == 95)
return CONFIG_CN_SPEECH_COMMAND_ID95;
else if (i == 96)
return CONFIG_CN_SPEECH_COMMAND_ID96;
else if (i == 97)
return CONFIG_CN_SPEECH_COMMAND_ID97;
else if (i == 98)
return CONFIG_CN_SPEECH_COMMAND_ID98;
else if (i == 99)
return CONFIG_CN_SPEECH_COMMAND_ID99;
else if (i == 100)
return CONFIG_CN_SPEECH_COMMAND_ID100;
else if (i == 101)
return CONFIG_CN_SPEECH_COMMAND_ID101;
else if (i == 102)
return CONFIG_CN_SPEECH_COMMAND_ID102;
else if (i == 103)
return CONFIG_CN_SPEECH_COMMAND_ID103;
else if (i == 104)
return CONFIG_CN_SPEECH_COMMAND_ID104;
else if (i == 105)
return CONFIG_CN_SPEECH_COMMAND_ID105;
else if (i == 106)
return CONFIG_CN_SPEECH_COMMAND_ID106;
else if (i == 107)
return CONFIG_CN_SPEECH_COMMAND_ID107;
else if (i == 108)
return CONFIG_CN_SPEECH_COMMAND_ID108;
else if (i == 109)
return CONFIG_CN_SPEECH_COMMAND_ID109;
else if (i == 110)
return CONFIG_CN_SPEECH_COMMAND_ID110;
else if (i == 111)
return CONFIG_CN_SPEECH_COMMAND_ID111;
else if (i == 112)
return CONFIG_CN_SPEECH_COMMAND_ID112;
else if (i == 113)
return CONFIG_CN_SPEECH_COMMAND_ID113;
else if (i == 114)
return CONFIG_CN_SPEECH_COMMAND_ID114;
else if (i == 115)
return CONFIG_CN_SPEECH_COMMAND_ID115;
else if (i == 116)
return CONFIG_CN_SPEECH_COMMAND_ID116;
else if (i == 117)
return CONFIG_CN_SPEECH_COMMAND_ID117;
else if (i == 118)
return CONFIG_CN_SPEECH_COMMAND_ID118;
else if (i == 119)
return CONFIG_CN_SPEECH_COMMAND_ID119;
else if (i == 120)
return CONFIG_CN_SPEECH_COMMAND_ID120;
else if (i == 121)
return CONFIG_CN_SPEECH_COMMAND_ID121;
else if (i == 122)
return CONFIG_CN_SPEECH_COMMAND_ID122;
else if (i == 123)
return CONFIG_CN_SPEECH_COMMAND_ID123;
else if (i == 124)
return CONFIG_CN_SPEECH_COMMAND_ID124;
else if (i == 125)
return CONFIG_CN_SPEECH_COMMAND_ID125;
else if (i == 126)
return CONFIG_CN_SPEECH_COMMAND_ID126;
else if (i == 127)
return CONFIG_CN_SPEECH_COMMAND_ID127;
else if (i == 128)
return CONFIG_CN_SPEECH_COMMAND_ID128;
else if (i == 129)
return CONFIG_CN_SPEECH_COMMAND_ID129;
else if (i == 130)
return CONFIG_CN_SPEECH_COMMAND_ID130;
else if (i == 131)
return CONFIG_CN_SPEECH_COMMAND_ID131;
else if (i == 132)
return CONFIG_CN_SPEECH_COMMAND_ID132;
else if (i == 133)
return CONFIG_CN_SPEECH_COMMAND_ID133;
else if (i == 134)
return CONFIG_CN_SPEECH_COMMAND_ID134;
else if (i == 135)
return CONFIG_CN_SPEECH_COMMAND_ID135;
else if (i == 136)
return CONFIG_CN_SPEECH_COMMAND_ID136;
else if (i == 137)
return CONFIG_CN_SPEECH_COMMAND_ID137;
else if (i == 138)
return CONFIG_CN_SPEECH_COMMAND_ID138;
else if (i == 139)
return CONFIG_CN_SPEECH_COMMAND_ID139;
else if (i == 140)
return CONFIG_CN_SPEECH_COMMAND_ID140;
else if (i == 141)
return CONFIG_CN_SPEECH_COMMAND_ID141;
else if (i == 142)
return CONFIG_CN_SPEECH_COMMAND_ID142;
else if (i == 143)
return CONFIG_CN_SPEECH_COMMAND_ID143;
else if (i == 144)
return CONFIG_CN_SPEECH_COMMAND_ID144;
else if (i == 145)
return CONFIG_CN_SPEECH_COMMAND_ID145;
else if (i == 146)
return CONFIG_CN_SPEECH_COMMAND_ID146;
else if (i == 147)
return CONFIG_CN_SPEECH_COMMAND_ID147;
else if (i == 148)
return CONFIG_CN_SPEECH_COMMAND_ID148;
else if (i == 149)
return CONFIG_CN_SPEECH_COMMAND_ID149;
else if (i == 150)
return CONFIG_CN_SPEECH_COMMAND_ID150;
else if (i == 151)
return CONFIG_CN_SPEECH_COMMAND_ID151;
else if (i == 152)
return CONFIG_CN_SPEECH_COMMAND_ID152;
else if (i == 153)
return CONFIG_CN_SPEECH_COMMAND_ID153;
else if (i == 154)
return CONFIG_CN_SPEECH_COMMAND_ID154;
else if (i == 155)
return CONFIG_CN_SPEECH_COMMAND_ID155;
else if (i == 156)
return CONFIG_CN_SPEECH_COMMAND_ID156;
else if (i == 157)
return CONFIG_CN_SPEECH_COMMAND_ID157;
else if (i == 158)
return CONFIG_CN_SPEECH_COMMAND_ID158;
else if (i == 159)
return CONFIG_CN_SPEECH_COMMAND_ID159;
else if (i == 160)
return CONFIG_CN_SPEECH_COMMAND_ID160;
else if (i == 161)
return CONFIG_CN_SPEECH_COMMAND_ID161;
else if (i == 162)
return CONFIG_CN_SPEECH_COMMAND_ID162;
else if (i == 163)
return CONFIG_CN_SPEECH_COMMAND_ID163;
else if (i == 164)
return CONFIG_CN_SPEECH_COMMAND_ID164;
else if (i == 165)
return CONFIG_CN_SPEECH_COMMAND_ID165;
else if (i == 166)
return CONFIG_CN_SPEECH_COMMAND_ID166;
else if (i == 167)
return CONFIG_CN_SPEECH_COMMAND_ID167;
else if (i == 168)
return CONFIG_CN_SPEECH_COMMAND_ID168;
else if (i == 169)
return CONFIG_CN_SPEECH_COMMAND_ID169;
else if (i == 170)
return CONFIG_CN_SPEECH_COMMAND_ID170;
else if (i == 171)
return CONFIG_CN_SPEECH_COMMAND_ID171;
else if (i == 172)
return CONFIG_CN_SPEECH_COMMAND_ID172;
else if (i == 173)
return CONFIG_CN_SPEECH_COMMAND_ID173;
else if (i == 174)
return CONFIG_CN_SPEECH_COMMAND_ID174;
else if (i == 175)
return CONFIG_CN_SPEECH_COMMAND_ID175;
else if (i == 176)
return CONFIG_CN_SPEECH_COMMAND_ID176;
else if (i == 177)
return CONFIG_CN_SPEECH_COMMAND_ID177;
else if (i == 178)
return CONFIG_CN_SPEECH_COMMAND_ID178;
else if (i == 179)
return CONFIG_CN_SPEECH_COMMAND_ID179;
else if (i == 180)
return CONFIG_CN_SPEECH_COMMAND_ID180;
else if (i == 181)
return CONFIG_CN_SPEECH_COMMAND_ID181;
else if (i == 182)
return CONFIG_CN_SPEECH_COMMAND_ID182;
else if (i == 183)
return CONFIG_CN_SPEECH_COMMAND_ID183;
else if (i == 184)
return CONFIG_CN_SPEECH_COMMAND_ID184;
else if (i == 185)
return CONFIG_CN_SPEECH_COMMAND_ID185;
else if (i == 186)
return CONFIG_CN_SPEECH_COMMAND_ID186;
else if (i == 187)
return CONFIG_CN_SPEECH_COMMAND_ID187;
else if (i == 188)
return CONFIG_CN_SPEECH_COMMAND_ID188;
else if (i == 189)
return CONFIG_CN_SPEECH_COMMAND_ID189;
else if (i == 190)
return CONFIG_CN_SPEECH_COMMAND_ID190;
else if (i == 191)
return CONFIG_CN_SPEECH_COMMAND_ID191;
else if (i == 192)
return CONFIG_CN_SPEECH_COMMAND_ID192;
else if (i == 193)
return CONFIG_CN_SPEECH_COMMAND_ID193;
else if (i == 194)
return CONFIG_CN_SPEECH_COMMAND_ID194;
else if (i == 195)
return CONFIG_CN_SPEECH_COMMAND_ID195;
else if (i == 196)
return CONFIG_CN_SPEECH_COMMAND_ID196;
else if (i == 197)
return CONFIG_CN_SPEECH_COMMAND_ID197;
else if (i == 198)
return CONFIG_CN_SPEECH_COMMAND_ID198;
else if (i == 199)
return CONFIG_CN_SPEECH_COMMAND_ID199;
else
return NULL;
#else
return NULL;
#endif
#else
return NULL;
#endif
}
char *get_id_name_en(int i)
{
#if defined CONFIG_USE_MULTINET && defined CONFIG_SR_MN_EN_MULTINET5_SINGLE_RECOGNITION_QUANT8
if (i == 0)
return CONFIG_EN_SPEECH_COMMAND_ID0;
else if (i == 1)
return CONFIG_EN_SPEECH_COMMAND_ID1;
else if (i == 2)
return CONFIG_EN_SPEECH_COMMAND_ID2;
else if (i == 3)
return CONFIG_EN_SPEECH_COMMAND_ID3;
else if (i == 4)
return CONFIG_EN_SPEECH_COMMAND_ID4;
else if (i == 5)
return CONFIG_EN_SPEECH_COMMAND_ID5;
else if (i == 6)
return CONFIG_EN_SPEECH_COMMAND_ID6;
else if (i == 7)
return CONFIG_EN_SPEECH_COMMAND_ID7;
else if (i == 8)
return CONFIG_EN_SPEECH_COMMAND_ID8;
else if (i == 9)
return CONFIG_EN_SPEECH_COMMAND_ID9;
else if (i == 10)
return CONFIG_EN_SPEECH_COMMAND_ID10;
else if (i == 11)
return CONFIG_EN_SPEECH_COMMAND_ID11;
else if (i == 12)
return CONFIG_EN_SPEECH_COMMAND_ID12;
else if (i == 13)
return CONFIG_EN_SPEECH_COMMAND_ID13;
else if (i == 14)
return CONFIG_EN_SPEECH_COMMAND_ID14;
else if (i == 15)
return CONFIG_EN_SPEECH_COMMAND_ID15;
else if (i == 16)
return CONFIG_EN_SPEECH_COMMAND_ID16;
else if (i == 17)
return CONFIG_EN_SPEECH_COMMAND_ID17;
else if (i == 18)
return CONFIG_EN_SPEECH_COMMAND_ID18;
else if (i == 19)
return CONFIG_EN_SPEECH_COMMAND_ID19;
else if (i == 20)
return CONFIG_EN_SPEECH_COMMAND_ID20;
else if (i == 21)
return CONFIG_EN_SPEECH_COMMAND_ID21;
else if (i == 22)
return CONFIG_EN_SPEECH_COMMAND_ID22;
else if (i == 23)
return CONFIG_EN_SPEECH_COMMAND_ID23;
else if (i == 24)
return CONFIG_EN_SPEECH_COMMAND_ID24;
else if (i == 25)
return CONFIG_EN_SPEECH_COMMAND_ID25;
else if (i == 26)
return CONFIG_EN_SPEECH_COMMAND_ID26;
else if (i == 27)
return CONFIG_EN_SPEECH_COMMAND_ID27;
else if (i == 28)
return CONFIG_EN_SPEECH_COMMAND_ID28;
else if (i == 29)
return CONFIG_EN_SPEECH_COMMAND_ID29;
else if (i == 30)
return CONFIG_EN_SPEECH_COMMAND_ID30;
else if (i == 31)
return CONFIG_EN_SPEECH_COMMAND_ID31;
else if (i == 32)
return CONFIG_EN_SPEECH_COMMAND_ID32;
else if (i == 33)
return CONFIG_EN_SPEECH_COMMAND_ID33;
else if (i == 34)
return CONFIG_EN_SPEECH_COMMAND_ID34;
else if (i == 35)
return CONFIG_EN_SPEECH_COMMAND_ID35;
else if (i == 36)
return CONFIG_EN_SPEECH_COMMAND_ID36;
else if (i == 37)
return CONFIG_EN_SPEECH_COMMAND_ID37;
else if (i == 38)
return CONFIG_EN_SPEECH_COMMAND_ID38;
else if (i == 39)
return CONFIG_EN_SPEECH_COMMAND_ID39;
else if (i == 40)
return CONFIG_EN_SPEECH_COMMAND_ID40;
else if (i == 41)
return CONFIG_EN_SPEECH_COMMAND_ID41;
else if (i == 42)
return CONFIG_EN_SPEECH_COMMAND_ID42;
else if (i == 43)
return CONFIG_EN_SPEECH_COMMAND_ID43;
else if (i == 44)
return CONFIG_EN_SPEECH_COMMAND_ID44;
else if (i == 45)
return CONFIG_EN_SPEECH_COMMAND_ID45;
else if (i == 46)
return CONFIG_EN_SPEECH_COMMAND_ID46;
else if (i == 47)
return CONFIG_EN_SPEECH_COMMAND_ID47;
else if (i == 48)
return CONFIG_EN_SPEECH_COMMAND_ID48;
else if (i == 49)
return CONFIG_EN_SPEECH_COMMAND_ID49;
else if (i == 50)
return CONFIG_EN_SPEECH_COMMAND_ID50;
else if (i == 51)
return CONFIG_EN_SPEECH_COMMAND_ID51;
else if (i == 52)
return CONFIG_EN_SPEECH_COMMAND_ID52;
else if (i == 53)
return CONFIG_EN_SPEECH_COMMAND_ID53;
else if (i == 54)
return CONFIG_EN_SPEECH_COMMAND_ID54;
else if (i == 55)
return CONFIG_EN_SPEECH_COMMAND_ID55;
else if (i == 56)
return CONFIG_EN_SPEECH_COMMAND_ID56;
else if (i == 57)
return CONFIG_EN_SPEECH_COMMAND_ID57;
else if (i == 58)
return CONFIG_EN_SPEECH_COMMAND_ID58;
else if (i == 59)
return CONFIG_EN_SPEECH_COMMAND_ID59;
else if (i == 60)
return CONFIG_EN_SPEECH_COMMAND_ID60;
else if (i == 61)
return CONFIG_EN_SPEECH_COMMAND_ID61;
else if (i == 62)
return CONFIG_EN_SPEECH_COMMAND_ID62;
else if (i == 63)
return CONFIG_EN_SPEECH_COMMAND_ID63;
else if (i == 64)
return CONFIG_EN_SPEECH_COMMAND_ID64;
else if (i == 65)
return CONFIG_EN_SPEECH_COMMAND_ID65;
else if (i == 66)
return CONFIG_EN_SPEECH_COMMAND_ID66;
else if (i == 67)
return CONFIG_EN_SPEECH_COMMAND_ID67;
else if (i == 68)
return CONFIG_EN_SPEECH_COMMAND_ID68;
else if (i == 69)
return CONFIG_EN_SPEECH_COMMAND_ID69;
else if (i == 70)
return CONFIG_EN_SPEECH_COMMAND_ID70;
else if (i == 71)
return CONFIG_EN_SPEECH_COMMAND_ID71;
else if (i == 72)
return CONFIG_EN_SPEECH_COMMAND_ID72;
else if (i == 73)
return CONFIG_EN_SPEECH_COMMAND_ID73;
else if (i == 74)
return CONFIG_EN_SPEECH_COMMAND_ID74;
else if (i == 75)
return CONFIG_EN_SPEECH_COMMAND_ID75;
else if (i == 76)
return CONFIG_EN_SPEECH_COMMAND_ID76;
else if (i == 77)
return CONFIG_EN_SPEECH_COMMAND_ID77;
else if (i == 78)
return CONFIG_EN_SPEECH_COMMAND_ID78;
else if (i == 79)
return CONFIG_EN_SPEECH_COMMAND_ID79;
else if (i == 80)
return CONFIG_EN_SPEECH_COMMAND_ID80;
else if (i == 81)
return CONFIG_EN_SPEECH_COMMAND_ID81;
else if (i == 82)
return CONFIG_EN_SPEECH_COMMAND_ID82;
else if (i == 83)
return CONFIG_EN_SPEECH_COMMAND_ID83;
else if (i == 84)
return CONFIG_EN_SPEECH_COMMAND_ID84;
else if (i == 85)
return CONFIG_EN_SPEECH_COMMAND_ID85;
else if (i == 86)
return CONFIG_EN_SPEECH_COMMAND_ID86;
else if (i == 87)
return CONFIG_EN_SPEECH_COMMAND_ID87;
else if (i == 88)
return CONFIG_EN_SPEECH_COMMAND_ID88;
else if (i == 89)
return CONFIG_EN_SPEECH_COMMAND_ID89;
else if (i == 90)
return CONFIG_EN_SPEECH_COMMAND_ID90;
else if (i == 91)
return CONFIG_EN_SPEECH_COMMAND_ID91;
else if (i == 92)
return CONFIG_EN_SPEECH_COMMAND_ID92;
else if (i == 93)
return CONFIG_EN_SPEECH_COMMAND_ID93;
else if (i == 94)
return CONFIG_EN_SPEECH_COMMAND_ID94;
else if (i == 95)
return CONFIG_EN_SPEECH_COMMAND_ID95;
else if (i == 96)
return CONFIG_EN_SPEECH_COMMAND_ID96;
else if (i == 97)
return CONFIG_EN_SPEECH_COMMAND_ID97;
else if (i == 98)
return CONFIG_EN_SPEECH_COMMAND_ID98;
else if (i == 99)
return CONFIG_EN_SPEECH_COMMAND_ID99;
else if (i == 100)
return CONFIG_EN_SPEECH_COMMAND_ID100;
else if (i == 101)
return CONFIG_EN_SPEECH_COMMAND_ID101;
else if (i == 102)
return CONFIG_EN_SPEECH_COMMAND_ID102;
else if (i == 103)
return CONFIG_EN_SPEECH_COMMAND_ID103;
else if (i == 104)
return CONFIG_EN_SPEECH_COMMAND_ID104;
else if (i == 105)
return CONFIG_EN_SPEECH_COMMAND_ID105;
else if (i == 106)
return CONFIG_EN_SPEECH_COMMAND_ID106;
else if (i == 107)
return CONFIG_EN_SPEECH_COMMAND_ID107;
else if (i == 108)
return CONFIG_EN_SPEECH_COMMAND_ID108;
else if (i == 109)
return CONFIG_EN_SPEECH_COMMAND_ID109;
else if (i == 110)
return CONFIG_EN_SPEECH_COMMAND_ID110;
else if (i == 111)
return CONFIG_EN_SPEECH_COMMAND_ID111;
else if (i == 112)
return CONFIG_EN_SPEECH_COMMAND_ID112;
else if (i == 113)
return CONFIG_EN_SPEECH_COMMAND_ID113;
else if (i == 114)
return CONFIG_EN_SPEECH_COMMAND_ID114;
else if (i == 115)
return CONFIG_EN_SPEECH_COMMAND_ID115;
else if (i == 116)
return CONFIG_EN_SPEECH_COMMAND_ID116;
else if (i == 117)
return CONFIG_EN_SPEECH_COMMAND_ID117;
else if (i == 118)
return CONFIG_EN_SPEECH_COMMAND_ID118;
else if (i == 119)
return CONFIG_EN_SPEECH_COMMAND_ID119;
else if (i == 120)
return CONFIG_EN_SPEECH_COMMAND_ID120;
else if (i == 121)
return CONFIG_EN_SPEECH_COMMAND_ID121;
else if (i == 122)
return CONFIG_EN_SPEECH_COMMAND_ID122;
else if (i == 123)
return CONFIG_EN_SPEECH_COMMAND_ID123;
else if (i == 124)
return CONFIG_EN_SPEECH_COMMAND_ID124;
else if (i == 125)
return CONFIG_EN_SPEECH_COMMAND_ID125;
else if (i == 126)
return CONFIG_EN_SPEECH_COMMAND_ID126;
else if (i == 127)
return CONFIG_EN_SPEECH_COMMAND_ID127;
else if (i == 128)
return CONFIG_EN_SPEECH_COMMAND_ID128;
else if (i == 129)
return CONFIG_EN_SPEECH_COMMAND_ID129;
else if (i == 130)
return CONFIG_EN_SPEECH_COMMAND_ID130;
else if (i == 131)
return CONFIG_EN_SPEECH_COMMAND_ID131;
else if (i == 132)
return CONFIG_EN_SPEECH_COMMAND_ID132;
else if (i == 133)
return CONFIG_EN_SPEECH_COMMAND_ID133;
else if (i == 134)
return CONFIG_EN_SPEECH_COMMAND_ID134;
else if (i == 135)
return CONFIG_EN_SPEECH_COMMAND_ID135;
else if (i == 136)
return CONFIG_EN_SPEECH_COMMAND_ID136;
else if (i == 137)
return CONFIG_EN_SPEECH_COMMAND_ID137;
else if (i == 138)
return CONFIG_EN_SPEECH_COMMAND_ID138;
else if (i == 139)
return CONFIG_EN_SPEECH_COMMAND_ID139;
else if (i == 140)
return CONFIG_EN_SPEECH_COMMAND_ID140;
else if (i == 141)
return CONFIG_EN_SPEECH_COMMAND_ID141;
else if (i == 142)
return CONFIG_EN_SPEECH_COMMAND_ID142;
else if (i == 143)
return CONFIG_EN_SPEECH_COMMAND_ID143;
else if (i == 144)
return CONFIG_EN_SPEECH_COMMAND_ID144;
else if (i == 145)
return CONFIG_EN_SPEECH_COMMAND_ID145;
else if (i == 146)
return CONFIG_EN_SPEECH_COMMAND_ID146;
else if (i == 147)
return CONFIG_EN_SPEECH_COMMAND_ID147;
else if (i == 148)
return CONFIG_EN_SPEECH_COMMAND_ID148;
else if (i == 149)
return CONFIG_EN_SPEECH_COMMAND_ID149;
else if (i == 150)
return CONFIG_EN_SPEECH_COMMAND_ID150;
else if (i == 151)
return CONFIG_EN_SPEECH_COMMAND_ID151;
else if (i == 152)
return CONFIG_EN_SPEECH_COMMAND_ID152;
else if (i == 153)
return CONFIG_EN_SPEECH_COMMAND_ID153;
else if (i == 154)
return CONFIG_EN_SPEECH_COMMAND_ID154;
else if (i == 155)
return CONFIG_EN_SPEECH_COMMAND_ID155;
else if (i == 156)
return CONFIG_EN_SPEECH_COMMAND_ID156;
else if (i == 157)
return CONFIG_EN_SPEECH_COMMAND_ID157;
else if (i == 158)
return CONFIG_EN_SPEECH_COMMAND_ID158;
else if (i == 159)
return CONFIG_EN_SPEECH_COMMAND_ID159;
else if (i == 160)
return CONFIG_EN_SPEECH_COMMAND_ID160;
else if (i == 161)
return CONFIG_EN_SPEECH_COMMAND_ID161;
else if (i == 162)
return CONFIG_EN_SPEECH_COMMAND_ID162;
else if (i == 163)
return CONFIG_EN_SPEECH_COMMAND_ID163;
else if (i == 164)
return CONFIG_EN_SPEECH_COMMAND_ID164;
else if (i == 165)
return CONFIG_EN_SPEECH_COMMAND_ID165;
else if (i == 166)
return CONFIG_EN_SPEECH_COMMAND_ID166;
else if (i == 167)
return CONFIG_EN_SPEECH_COMMAND_ID167;
else if (i == 168)
return CONFIG_EN_SPEECH_COMMAND_ID168;
else if (i == 169)
return CONFIG_EN_SPEECH_COMMAND_ID169;
else if (i == 170)
return CONFIG_EN_SPEECH_COMMAND_ID170;
else if (i == 171)
return CONFIG_EN_SPEECH_COMMAND_ID171;
else if (i == 172)
return CONFIG_EN_SPEECH_COMMAND_ID172;
else if (i == 173)
return CONFIG_EN_SPEECH_COMMAND_ID173;
else if (i == 174)
return CONFIG_EN_SPEECH_COMMAND_ID174;
else if (i == 175)
return CONFIG_EN_SPEECH_COMMAND_ID175;
else if (i == 176)
return CONFIG_EN_SPEECH_COMMAND_ID176;
else if (i == 177)
return CONFIG_EN_SPEECH_COMMAND_ID177;
else if (i == 178)
return CONFIG_EN_SPEECH_COMMAND_ID178;
else if (i == 179)
return CONFIG_EN_SPEECH_COMMAND_ID179;
else if (i == 180)
return CONFIG_EN_SPEECH_COMMAND_ID180;
else if (i == 181)
return CONFIG_EN_SPEECH_COMMAND_ID181;
else if (i == 182)
return CONFIG_EN_SPEECH_COMMAND_ID182;
else if (i == 183)
return CONFIG_EN_SPEECH_COMMAND_ID183;
else if (i == 184)
return CONFIG_EN_SPEECH_COMMAND_ID184;
else if (i == 185)
return CONFIG_EN_SPEECH_COMMAND_ID185;
else if (i == 186)
return CONFIG_EN_SPEECH_COMMAND_ID186;
else if (i == 187)
return CONFIG_EN_SPEECH_COMMAND_ID187;
else if (i == 188)
return CONFIG_EN_SPEECH_COMMAND_ID188;
else if (i == 189)
return CONFIG_EN_SPEECH_COMMAND_ID189;
else if (i == 190)
return CONFIG_EN_SPEECH_COMMAND_ID190;
else if (i == 191)
return CONFIG_EN_SPEECH_COMMAND_ID191;
else if (i == 192)
return CONFIG_EN_SPEECH_COMMAND_ID192;
else if (i == 193)
return CONFIG_EN_SPEECH_COMMAND_ID193;
else if (i == 194)
return CONFIG_EN_SPEECH_COMMAND_ID194;
else if (i == 195)
return CONFIG_EN_SPEECH_COMMAND_ID195;
else if (i == 196)
return CONFIG_EN_SPEECH_COMMAND_ID196;
else if (i == 197)
return CONFIG_EN_SPEECH_COMMAND_ID197;
else if (i == 198)
return CONFIG_EN_SPEECH_COMMAND_ID198;
else if (i == 199)
return CONFIG_EN_SPEECH_COMMAND_ID199;
else
return NULL;
#else
return NULL;
#endif
}
esp_err_t esp_mn_commands_update_from_sdkconfig(esp_mn_iface_t *multinet, const model_iface_data_t *model_data)
{
esp_mn_commands_alloc();
int total_phrase_num = 0;
for (int i = 0; i < ESP_MN_MAX_PHRASE_NUM; i++) {
char *command_str = NULL;
int command_str_len = 0;
#if defined CONFIG_SR_MN_CN_MULTINET2_SINGLE_RECOGNITION || defined CONFIG_SR_MN_CN_MULTINET4_SINGLE_RECOGNITION || defined CONFIG_SR_MN_CN_MULTINET4_5_SINGLE_RECOGNITION || defined CONFIG_SR_MN_CN_MULTINET4_5_SINGLE_RECOGNITION_QUANT8
command_str = get_id_name_cn(i);
#elif defined CONFIG_SR_MN_EN_MULTINET5_SINGLE_RECOGNITION || defined CONFIG_SR_MN_EN_MULTINET5_SINGLE_RECOGNITION_QUANT8
command_str = get_id_name_en(i);
#else
ESP_LOGE(TAG, "Incorrect language");
#endif
command_str_len = strlen(command_str);
if (command_str_len <= 0) {
continue;
}
char *command_str_temp = calloc(command_str_len + 1, 1);
memcpy(command_str_temp, command_str, command_str_len);
char *token = strtok(command_str_temp, ",");
while (token != NULL) {
int len = strlen(token);
if (total_phrase_num > ESP_MN_MAX_PHRASE_NUM) {
ESP_LOGE(TAG, "The number of speech commands phrase must less than ESP_MN_MAX_PHRASE_NUM");
free(command_str_temp);
goto end;
}
esp_mn_commands_add(i, token);
total_phrase_num++;
token = strtok(NULL, ",");
}
free(command_str_temp);
}
end:
esp_mn_commands_print();
return esp_mn_commands_update(multinet, model_data);
}

187
src/model_path.c
View File

@ -1,6 +1,12 @@
#include "stdio.h" #include "stdio.h"
#include "sdkconfig.h" #include "sdkconfig.h"
#include "esp_spiffs.h" #include "esp_spiffs.h"
#include <sys/stat.h>
#include <sys/dirent.h>
#include "string.h"
#include "model_path.h"
#include "esp_wn_models.h"
#include "esp_mn_models.h"
char *get_model_base_path(void) char *get_model_base_path(void)
{ {
@ -13,9 +19,60 @@ char *get_model_base_path(void)
#endif #endif
} }
void srmodel_spiffs_init(void) srmodel_list_t *read_models_form_spiffs(const char *path)
{ {
printf("Initializing SPIFFS\n"); struct dirent *ret;
DIR *dir;
dir = opendir(path);
srmodel_list_t *models = malloc(sizeof(srmodel_list_t*));
models->num = 0;
int idx = 0;
if (dir != NULL)
{
while ((ret = readdir(dir)) != NULL)
{ // NULL if reach the end of directory
if (ret->d_type == DT_DIR) // continue if d_type is not file
continue;
int len = strlen(ret->d_name);
char *suffix = ret->d_name + len - 12;
if (strcmp(suffix, "_MODEL_INFO_") == 0)
models->num ++;
}
// allocate model names
models->model_name = malloc(models->num*sizeof(char*));
for (int i=0; i<models->num; i++) {
models->model_name[i] = (char*) calloc(MODEL_NAME_MAX_LENGTH, sizeof(char));
}
dir = opendir(path);
while ((ret = readdir(dir)) != NULL)
{ // NULL if reach the end of directory
if (ret->d_type == DT_DIR) // continue if d_type is not file
continue;
int len = strlen(ret->d_name);
char *suffix = ret->d_name + len - 12;
if (strcmp(suffix, "_MODEL_INFO_") == 0)
{
memcpy(models->model_name[idx], ret->d_name, (len-13)*sizeof(char));
// models->model_name[idx][len-13] = '\0';
idx ++;
}
}
}
return models;
}
srmodel_list_t* srmodel_spiffs_init(void)
{
printf("Initializing models from SPIFFS\n");
esp_vfs_spiffs_conf_t conf = { esp_vfs_spiffs_conf_t conf = {
.base_path = "/srmodel", .base_path = "/srmodel",
@ -36,7 +93,7 @@ void srmodel_spiffs_init(void)
} else { } else {
printf("Failed to initialize SPIFFS (%s)\n", esp_err_to_name(ret)); printf("Failed to initialize SPIFFS (%s)\n", esp_err_to_name(ret));
} }
return; return NULL;
} }
size_t total = 0, used = 0; size_t total = 0, used = 0;
@ -46,4 +103,128 @@ void srmodel_spiffs_init(void)
} else { } else {
printf("Partition size: total: %d, used: %d\n", total, used); printf("Partition size: total: %d, used: %d\n", total, used);
} }
// Read all model from path
return read_models_form_spiffs(conf.base_path);
} }
void srmodel_spiffs_deinit(srmodel_list_t *models)
{
if (models != NULL) {
if (models->num>0) {
for (int i=0; i<models->num; i++) {
free(models->model_name[i]);
}
}
free(models);
}
esp_vfs_spiffs_conf_t conf = {
.base_path = "/srmodel",
.partition_label = "model",
.max_files = 5,
.format_if_mount_failed = true
};
esp_vfs_spiffs_unregister(conf.partition_label);
}
srmodel_list_t *srmodel_config_init()
{
srmodel_list_t *models = malloc(sizeof(srmodel_list_t*));
models->num = 2;
models->model_name = malloc(models->num*sizeof(char*));
for (int i=0; i<models->num; i++) {
models->model_name[i] = (char*) calloc(MODEL_NAME_MAX_LENGTH, sizeof(char));
}
// If wakenet is selected in menuconfig, load the wakenet model
if (strcmp(WAKENET_MODEL_NAME, "NULL") == 0) {
models->num --;
free(models->model_name[models->num]);
} else {
strcpy(models->model_name[0], WAKENET_MODEL_NAME);
}
// If multinet is selected in menuconfig, load the multinet model
if (strcmp(MULTINET_MODEL_NAME, "NULL") == 0) {
models->num --;
free(models->model_name[models->num]);
} else {
strcpy(models->model_name[models->num-1], MULTINET_MODEL_NAME);
}
// could not find any avaliable models, return NULL
if (models->num == 0) {
free(models);
models = NULL;
}
return models;
}
void srmodel_config_deinit(srmodel_list_t *models)
{
if (models != NULL) {
if (models->num>0) {
for (int i=0; i<models->num; i++) {
free(models->model_name[i]);
}
}
free(models);
}
}
srmodel_list_t* esp_srmodel_init()
{
#ifdef CONFIG_IDF_TARGET_ESP32
return srmodel_config_init();
#else
return srmodel_spiffs_init();
#endif
}
void esp_srmodel_deinit(srmodel_list_t *models)
{
#ifdef CONFIG_IDF_TARGET_ESP32
return srmodel_config_deinit(models);
#else
return srmodel_spiffs_deinit(models);
#endif
}
// repackage strstr function to support needle==NULL
char *_esp_strstr_(const char *haystack, const char *needle)
{
if (needle == NULL) return haystack;
else return strstr(haystack, needle);
}
char *esp_srmodel_filter(srmodel_list_t *models, const char *keyword1, const char *keyword2)
{
if (models == NULL) return NULL;
// return the first model name including specific keyword
for (int i=0; i<models->num; i++) {
if (_esp_strstr_(models->model_name[i], keyword1) != NULL) {
if (_esp_strstr_(models->model_name[i], keyword2) != NULL)
return models->model_name[i];
}
}
return NULL;
}
int esp_srmodel_exists(srmodel_list_t *models, char *model_name) {
if (models == NULL) return -1;
for (int i=0; i<models->num; i++) {
if (strcmp(models->model_name[i], model_name) == 0) {
return i;
}
}
return -1;
}