EngineX-Iluvatar/enginex-mr_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Add inverse text normalization for online ASR (#1020)

Browse Source
This commit is contained in:
Fangjun Kuang
2024-06-17 18:39:23 +08:00
committed by GitHub
parent 6e09933d99
commit 349d957da2
12 changed files with 390 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
sherpa-onnx/python/csrc/online-recognizer.cc
View File

@@ -54,19 +54,20 @@ static void PybindOnlineRecognizerResult(py::module *m) {
static void PybindOnlineRecognizerConfig(py::module *m) {
using PyClass = OnlineRecognizerConfig;
py::class_<PyClass>(*m, "OnlineRecognizerConfig")
.def(
py::init<const FeatureExtractorConfig &, const OnlineModelConfig &,
const OnlineLMConfig &, const EndpointConfig &,
const OnlineCtcFstDecoderConfig &, bool, const std::string &,
int32_t, const std::string &, float, float, float>(),
py::arg("feat_config"), py::arg("model_config"),
py::arg("lm_config") = OnlineLMConfig(),
py::arg("endpoint_config") = EndpointConfig(),
py::arg("ctc_fst_decoder_config") = OnlineCtcFstDecoderConfig(),
py::arg("enable_endpoint"), py::arg("decoding_method"),
py::arg("max_active_paths") = 4, py::arg("hotwords_file") = "",
py::arg("hotwords_score") = 0, py::arg("blank_penalty") = 0.0,
py::arg("temperature_scale") = 2.0)
.def(py::init<const FeatureExtractorConfig &, const OnlineModelConfig &,
const OnlineLMConfig &, const EndpointConfig &,
const OnlineCtcFstDecoderConfig &, bool,
const std::string &, int32_t, const std::string &, float,
float, float, const std::string &, const std::string &>(),
py::arg("feat_config"), py::arg("model_config"),
py::arg("lm_config") = OnlineLMConfig(),
py::arg("endpoint_config") = EndpointConfig(),
py::arg("ctc_fst_decoder_config") = OnlineCtcFstDecoderConfig(),
py::arg("enable_endpoint"), py::arg("decoding_method"),
py::arg("max_active_paths") = 4, py::arg("hotwords_file") = "",
py::arg("hotwords_score") = 0, py::arg("blank_penalty") = 0.0,
py::arg("temperature_scale") = 2.0, py::arg("rule_fsts") = "",
py::arg("rule_fars") = "")
.def_readwrite("feat_config", &PyClass::feat_config)
.def_readwrite("model_config", &PyClass::model_config)
.def_readwrite("lm_config", &PyClass::lm_config)
@@ -79,6 +80,8 @@ static void PybindOnlineRecognizerConfig(py::module *m) {
.def_readwrite("hotwords_score", &PyClass::hotwords_score)
.def_readwrite("blank_penalty", &PyClass::blank_penalty)
.def_readwrite("temperature_scale", &PyClass::temperature_scale)
.def_readwrite("rule_fsts", &PyClass::rule_fsts)
.def_readwrite("rule_fars", &PyClass::rule_fars)
.def("__str__", &PyClass::ToString);
}

50
sherpa-onnx/python/sherpa_onnx/online_recognizer.py
View File

@@ -64,6 +64,8 @@ class OnlineRecognizer(object):
lm_scale: float = 0.1,
temperature_scale: float = 2.0,
debug: bool = False,
rule_fsts: str = "",
rule_fars: str = "",
):
"""
Please refer to
@@ -148,6 +150,12 @@ class OnlineRecognizer(object):
the log probability, you can get it from the directory where
your bpe model is generated. Only used when hotwords provided
and the modeling unit is bpe or cjkchar+bpe.
rule_fsts:
If not empty, it specifies fsts for inverse text normalization.
If there are multiple fsts, they are separated by a comma.
rule_fars:
If not empty, it specifies fst archives for inverse text normalization.
If there are multiple archives, they are separated by a comma.
"""
self = cls.__new__(cls)
_assert_file_exists(tokens)
@@ -217,6 +225,8 @@ class OnlineRecognizer(object):
hotwords_file=hotwords_file,
blank_penalty=blank_penalty,
temperature_scale=temperature_scale,
rule_fsts=rule_fsts,
rule_fars=rule_fars,
)
self.recognizer = _Recognizer(recognizer_config)
@@ -239,6 +249,8 @@ class OnlineRecognizer(object):
decoding_method: str = "greedy_search",
provider: str = "cpu",
debug: bool = False,
rule_fsts: str = "",
rule_fars: str = "",
):
"""
Please refer to
@@ -283,6 +295,12 @@ class OnlineRecognizer(object):
The only valid value is greedy_search.
provider:
onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
rule_fsts:
If not empty, it specifies fsts for inverse text normalization.
If there are multiple fsts, they are separated by a comma.
rule_fars:
If not empty, it specifies fst archives for inverse text normalization.
If there are multiple archives, they are separated by a comma.
"""
self = cls.__new__(cls)
_assert_file_exists(tokens)
@@ -322,6 +340,8 @@ class OnlineRecognizer(object):
endpoint_config=endpoint_config,
enable_endpoint=enable_endpoint_detection,
decoding_method=decoding_method,
rule_fsts=rule_fsts,
rule_fars=rule_fars,
)
self.recognizer = _Recognizer(recognizer_config)
@@ -345,6 +365,8 @@ class OnlineRecognizer(object):
ctc_max_active: int = 3000,
provider: str = "cpu",
debug: bool = False,
rule_fsts: str = "",
rule_fars: str = "",
):
"""
Please refer to
@@ -393,6 +415,12 @@ class OnlineRecognizer(object):
active paths at a time.
provider:
onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
rule_fsts:
If not empty, it specifies fsts for inverse text normalization.
If there are multiple fsts, they are separated by a comma.
rule_fars:
If not empty, it specifies fst archives for inverse text normalization.
If there are multiple archives, they are separated by a comma.
"""
self = cls.__new__(cls)
_assert_file_exists(tokens)
@@ -433,6 +461,8 @@ class OnlineRecognizer(object):
ctc_fst_decoder_config=ctc_fst_decoder_config,
enable_endpoint=enable_endpoint_detection,
decoding_method=decoding_method,
rule_fsts=rule_fsts,
rule_fars=rule_fars,
)
self.recognizer = _Recognizer(recognizer_config)
@@ -454,6 +484,8 @@ class OnlineRecognizer(object):
decoding_method: str = "greedy_search",
provider: str = "cpu",
debug: bool = False,
rule_fsts: str = "",
rule_fars: str = "",
):
"""
Please refer to
@@ -497,6 +529,12 @@ class OnlineRecognizer(object):
onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
debug:
True to show meta data in the model.
rule_fsts:
If not empty, it specifies fsts for inverse text normalization.
If there are multiple fsts, they are separated by a comma.
rule_fars:
If not empty, it specifies fst archives for inverse text normalization.
If there are multiple archives, they are separated by a comma.
"""
self = cls.__new__(cls)
_assert_file_exists(tokens)
@@ -533,6 +571,8 @@ class OnlineRecognizer(object):
endpoint_config=endpoint_config,
enable_endpoint=enable_endpoint_detection,
decoding_method=decoding_method,
rule_fsts=rule_fsts,
rule_fars=rule_fars,
)
self.recognizer = _Recognizer(recognizer_config)
@@ -556,6 +596,8 @@ class OnlineRecognizer(object):
decoding_method: str = "greedy_search",
provider: str = "cpu",
debug: bool = False,
rule_fsts: str = "",
rule_fars: str = "",
):
"""
Please refer to
@@ -602,6 +644,12 @@ class OnlineRecognizer(object):
The only valid value is greedy_search.
provider:
onnxruntime execution providers. Valid values are: cpu, cuda, coreml.
rule_fsts:
If not empty, it specifies fsts for inverse text normalization.
If there are multiple fsts, they are separated by a comma.
rule_fars:
If not empty, it specifies fst archives for inverse text normalization.
If there are multiple archives, they are separated by a comma.
"""
self = cls.__new__(cls)
_assert_file_exists(tokens)
@@ -640,6 +688,8 @@ class OnlineRecognizer(object):
endpoint_config=endpoint_config,
enable_endpoint=enable_endpoint_detection,
decoding_method=decoding_method,
rule_fsts=rule_fsts,
rule_fars=rule_fars,
)
self.recognizer = _Recognizer(recognizer_config)