Made in Vancouver, Canada by Picovoice
Leopard is an on-device Speech-to-Text engine. Leopard is:
- offline and runs locally without an internet connection.
- highly-accurate [1].
- compact and computationally-efficient [1].
- cross-platform. Linux (x86_64), Mac (x86_64), Windows (x86_64), web browsers, Android, iOS, Raspberry Pi, and Beagle Bone are supported. Linux (x86_64) is available for personal and non-commercial use free of charge. Other platforms are only available under the commercial license.
- customizable. Allows adding new words and adapting to different contexts (Available only under the commercial license).
This repository is provided for personal & non-commercial use only. Refer to LICENSE for details. If you wish to use Leopard in a commercial product request access here.
Leopard is shipped as a dynamic library. The binary files for supported platforms are located under lib and header files are at include. Bindings are available at binding to facilitate usage from higher-level languages/platforms. Demo applications are at demo. Finally, resources is a placeholder for data used by various applications within the repository.
The demo transcribes a set of audio files provided as command line arguments. The demo has been tested using Python 3.6 on a machine running Ubuntu 18.04 (x86_64). Note that the audio files need to be single-channel, 16KHz, and 16-bit linearly-encoded. For more information about audio requirements refer to pv_leopard.h. The following transcribes the WAV file located in the resource directory.
python demo/python/leopard_demo.py --audio_paths resources/audio_samples/test.wavIn order to transcribe multiple files provide their paths as below.
python demo/python/leopard_demo.py --audio_paths PATH_TO_AUDIO_FILE_1 PATH_TO_AUDIO_FILE_2 PATH_TO_AUDIO_FILE_3This demo application accepts a list of WAV files as input and returns their transcripts. Note that the demo expects the
audio files to be WAV, 16KHz, and 16-bit linearly-encoded. It does not perform any verification to assure the
compatibility or correctness of the input audio files. Set the current working directory to the root of the repository.
The demo can be built using gcc as below.
gcc -I include/ -O3 demo/c/leopard_demo.c -ldl -o leopard_demoThe usage can be attained by
./leopard_demoThen it can be used as follows
./leopard_demo \
./lib/linux/x86_64/libpv_leopard.so \
./lib/common/acoustic_model.pv \
./lib/common/language_model.pv \
./resources/license/leopard_eval_linux.lic \
./resources/audio_samples/test.wavleopard.py provides a Python binding. Below is a quick demonstration of how to construct an instance of it.
library_path = ... # the file is available under lib/linux/x86_64/libpv_leopard.so
acoustic_model_path = ... # the file is available under lib/common/acoustic_model.pv
language_model_path = ... # the file is available under lib/common/language_model.pv
license_path = ... # the file is available under resources/license/leopard_eval_linux.lic
handle = Leopard(library_path, acoustic_model_path, language_model_path, license_path)When initialized, valid sample rate can be retrieved using handle.sample_rate. Additionally, Leopard accepts
single-channel 16-bit linearly-encoded audio.
audio = ... # audio data to be transcribed as a NumPy array
transcript = handle.process(audio)When finished, release the acquired resources.
handle.delete()Leopard is implemented in ANSI C and therefore can be directly linked to C applications. pv_leopard.h header file contains relevant information. An instance of Leopard object can be constructed as follows.
const char *acoustic_model_path = ... // the file is available under lib/common/acoustic_model.pv
const char *language_model_path = ... // the file is available under lib/common/language_model.pv
const char *license_path = ... // the file is available under resources/license/leopard_eval_linux.lic
pv_leopard_t *handle;
const pv_status_t status = pv_leopard_init(acoustic_model_path, language_model_path, license_path, &handle);
if (status != PV_STATUS_SUCCESS) {
// error handling logic
}Now the handle can be used to process audio. Leopard accepts single-channel 16-bit linearly-encoded audio. The sample
rate can be retrieved using pv_sample_rate().
const int16_t *pcm = ... // audio data to be transcribed
const int32_t num_samples = ... // number of samples to process
char *transcript;
const pv_status_t status = pv_leopard_process(handle, pcm, num_samples, &transcript);
if (status != PV_STATUS_SUCCESS) {
// error handling logic
}
// the caller is required to free the transcription buffer when done.
free(transcript);Finally, when done be sure to release resources acquired.
pv_leopard_delete(handle);- Initial release.