正在显示
6 个修改的文件
包含
49 行增加
和
5 行删除
| @@ -80,6 +80,7 @@ def get_binaries(): | @@ -80,6 +80,7 @@ def get_binaries(): | ||
| 80 | "kaldi-decoder-core.dll", | 80 | "kaldi-decoder-core.dll", |
| 81 | "kaldi-native-fbank-core.dll", | 81 | "kaldi-native-fbank-core.dll", |
| 82 | "onnxruntime.dll", | 82 | "onnxruntime.dll", |
| 83 | + "ssentencepiece_core.dll", | ||
| 83 | "piper_phonemize.dll", | 84 | "piper_phonemize.dll", |
| 84 | "sherpa-onnx-c-api.dll", | 85 | "sherpa-onnx-c-api.dll", |
| 85 | "sherpa-onnx-core.dll", | 86 | "sherpa-onnx-core.dll", |
| @@ -13,4 +13,4 @@ Cflags: -I"${includedir}" | @@ -13,4 +13,4 @@ Cflags: -I"${includedir}" | ||
| 13 | # Note: -lcargs is required only for the following file | 13 | # Note: -lcargs is required only for the following file |
| 14 | # https://github.com/k2-fsa/sherpa-onnx/blob/master/c-api-examples/decode-file-c-api.c | 14 | # https://github.com/k2-fsa/sherpa-onnx/blob/master/c-api-examples/decode-file-c-api.c |
| 15 | # We add it here so that users don't need to specify -lcargs when compiling decode-file-c-api.c | 15 | # We add it here so that users don't need to specify -lcargs when compiling decode-file-c-api.c |
| 16 | -Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fst -lkaldi-native-fbank-core -lonnxruntime -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_WITH_CARGS@ @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@ | 16 | +Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fst -lkaldi-native-fbank-core -lonnxruntime -lssentencepiece_core -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_WITH_CARGS@ @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@ |
| @@ -13,4 +13,4 @@ Cflags: -I"${includedir}" | @@ -13,4 +13,4 @@ Cflags: -I"${includedir}" | ||
| 13 | # Note: -lcargs is required only for the following file | 13 | # Note: -lcargs is required only for the following file |
| 14 | # https://github.com/k2-fsa/sherpa-onnx/blob/master/c-api-examples/decode-file-c-api.c | 14 | # https://github.com/k2-fsa/sherpa-onnx/blob/master/c-api-examples/decode-file-c-api.c |
| 15 | # We add it here so that users don't need to specify -lcargs when compiling decode-file-c-api.c | 15 | # We add it here so that users don't need to specify -lcargs when compiling decode-file-c-api.c |
| 16 | -Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fstfar -lsherpa-onnx-fst -lkaldi-native-fbank-core -lpiper_phonemize -lespeak-ng -lucd -lonnxruntime -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_WITH_CARGS@ @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@ | 16 | +Libs: -L"${libdir}" -lsherpa-onnx-c-api -lsherpa-onnx-core -lkaldi-decoder-core -lsherpa-onnx-kaldifst-core -lsherpa-onnx-fstfar -lsherpa-onnx-fst -lkaldi-native-fbank-core -lpiper_phonemize -lespeak-ng -lucd -lonnxruntime -lssentencepiece_core -Wl,-rpath,${libdir} @SHERPA_ONNX_PKG_WITH_CARGS@ @SHERPA_ONNX_PKG_CONFIG_EXTRA_LIBS@ |
| @@ -59,9 +59,9 @@ static OfflineRecognitionResult Convert( | @@ -59,9 +59,9 @@ static OfflineRecognitionResult Convert( | ||
| 59 | mergeable = false; | 59 | mergeable = false; |
| 60 | 60 | ||
| 61 | if (i > 0) { | 61 | if (i > 0) { |
| 62 | - const uint8_t *p = reinterpret_cast<const uint8_t *>( | ||
| 63 | - sym_table[src.tokens[i - 1]].c_str()); | ||
| 64 | - if (p[0] < 0x80) { | 62 | + const uint8_t p = reinterpret_cast<const uint8_t *>( |
| 63 | + sym_table[src.tokens[i - 1]].c_str())[0]; | ||
| 64 | + if (p < 0x80) { | ||
| 65 | // put a space between ascii and non-ascii | 65 | // put a space between ascii and non-ascii |
| 66 | text.append(" "); | 66 | text.append(" "); |
| 67 | } | 67 | } |
| @@ -13,6 +13,8 @@ public class OfflineModelConfig { | @@ -13,6 +13,8 @@ public class OfflineModelConfig { | ||
| 13 | private final String provider; | 13 | private final String provider; |
| 14 | 14 | ||
| 15 | private final String modelType; | 15 | private final String modelType; |
| 16 | + private final String modelingUnit; | ||
| 17 | + private final String bpeVocab; | ||
| 16 | 18 | ||
| 17 | private OfflineModelConfig(Builder builder) { | 19 | private OfflineModelConfig(Builder builder) { |
| 18 | this.transducer = builder.transducer; | 20 | this.transducer = builder.transducer; |
| @@ -24,6 +26,8 @@ public class OfflineModelConfig { | @@ -24,6 +26,8 @@ public class OfflineModelConfig { | ||
| 24 | this.debug = builder.debug; | 26 | this.debug = builder.debug; |
| 25 | this.provider = builder.provider; | 27 | this.provider = builder.provider; |
| 26 | this.modelType = builder.modelType; | 28 | this.modelType = builder.modelType; |
| 29 | + this.modelingUnit = builder.modelingUnit; | ||
| 30 | + this.bpeVocab = builder.bpeVocab; | ||
| 27 | } | 31 | } |
| 28 | 32 | ||
| 29 | public static Builder builder() { | 33 | public static Builder builder() { |
| @@ -62,6 +66,13 @@ public class OfflineModelConfig { | @@ -62,6 +66,13 @@ public class OfflineModelConfig { | ||
| 62 | return modelType; | 66 | return modelType; |
| 63 | } | 67 | } |
| 64 | 68 | ||
| 69 | + public String getModelingUnit() { | ||
| 70 | + return modelingUnit; | ||
| 71 | + } | ||
| 72 | + | ||
| 73 | + public String getBpeVocab() { | ||
| 74 | + return bpeVocab; | ||
| 75 | + } | ||
| 65 | 76 | ||
| 66 | public static class Builder { | 77 | public static class Builder { |
| 67 | private OfflineParaformerModelConfig paraformer = OfflineParaformerModelConfig.builder().build(); | 78 | private OfflineParaformerModelConfig paraformer = OfflineParaformerModelConfig.builder().build(); |
| @@ -73,6 +84,8 @@ public class OfflineModelConfig { | @@ -73,6 +84,8 @@ public class OfflineModelConfig { | ||
| 73 | private boolean debug = true; | 84 | private boolean debug = true; |
| 74 | private String provider = "cpu"; | 85 | private String provider = "cpu"; |
| 75 | private String modelType = ""; | 86 | private String modelType = ""; |
| 87 | + private String modelingUnit = "cjkchar"; | ||
| 88 | + private String bpeVocab = ""; | ||
| 76 | 89 | ||
| 77 | public OfflineModelConfig build() { | 90 | public OfflineModelConfig build() { |
| 78 | return new OfflineModelConfig(this); | 91 | return new OfflineModelConfig(this); |
| @@ -122,5 +135,13 @@ public class OfflineModelConfig { | @@ -122,5 +135,13 @@ public class OfflineModelConfig { | ||
| 122 | this.modelType = modelType; | 135 | this.modelType = modelType; |
| 123 | return this; | 136 | return this; |
| 124 | } | 137 | } |
| 138 | + | ||
| 139 | + public void setModelingUnit(String modelingUnit) { | ||
| 140 | + this.modelingUnit = modelingUnit; | ||
| 141 | + } | ||
| 142 | + | ||
| 143 | + public void setBpeVocab(String bpeVocab) { | ||
| 144 | + this.bpeVocab = bpeVocab; | ||
| 145 | + } | ||
| 125 | } | 146 | } |
| 126 | } | 147 | } |
| @@ -13,6 +13,8 @@ public class OnlineModelConfig { | @@ -13,6 +13,8 @@ public class OnlineModelConfig { | ||
| 13 | private final boolean debug; | 13 | private final boolean debug; |
| 14 | private final String provider; | 14 | private final String provider; |
| 15 | private final String modelType; | 15 | private final String modelType; |
| 16 | + private final String modelingUnit; | ||
| 17 | + private final String bpeVocab; | ||
| 16 | 18 | ||
| 17 | private OnlineModelConfig(Builder builder) { | 19 | private OnlineModelConfig(Builder builder) { |
| 18 | this.transducer = builder.transducer; | 20 | this.transducer = builder.transducer; |
| @@ -24,6 +26,8 @@ public class OnlineModelConfig { | @@ -24,6 +26,8 @@ public class OnlineModelConfig { | ||
| 24 | this.debug = builder.debug; | 26 | this.debug = builder.debug; |
| 25 | this.provider = builder.provider; | 27 | this.provider = builder.provider; |
| 26 | this.modelType = builder.modelType; | 28 | this.modelType = builder.modelType; |
| 29 | + this.modelingUnit = builder.modelingUnit; | ||
| 30 | + this.bpeVocab = builder.bpeVocab; | ||
| 27 | } | 31 | } |
| 28 | 32 | ||
| 29 | public static Builder builder() { | 33 | public static Builder builder() { |
| @@ -66,6 +70,14 @@ public class OnlineModelConfig { | @@ -66,6 +70,14 @@ public class OnlineModelConfig { | ||
| 66 | return modelType; | 70 | return modelType; |
| 67 | } | 71 | } |
| 68 | 72 | ||
| 73 | + public String getModelingUnit() { | ||
| 74 | + return modelingUnit; | ||
| 75 | + } | ||
| 76 | + | ||
| 77 | + public String getBpeVocab() { | ||
| 78 | + return bpeVocab; | ||
| 79 | + } | ||
| 80 | + | ||
| 69 | public static class Builder { | 81 | public static class Builder { |
| 70 | private OnlineParaformerModelConfig paraformer = OnlineParaformerModelConfig.builder().build(); | 82 | private OnlineParaformerModelConfig paraformer = OnlineParaformerModelConfig.builder().build(); |
| 71 | private OnlineTransducerModelConfig transducer = OnlineTransducerModelConfig.builder().build(); | 83 | private OnlineTransducerModelConfig transducer = OnlineTransducerModelConfig.builder().build(); |
| @@ -76,6 +88,8 @@ public class OnlineModelConfig { | @@ -76,6 +88,8 @@ public class OnlineModelConfig { | ||
| 76 | private boolean debug = true; | 88 | private boolean debug = true; |
| 77 | private String provider = "cpu"; | 89 | private String provider = "cpu"; |
| 78 | private String modelType = ""; | 90 | private String modelType = ""; |
| 91 | + private String modelingUnit = "cjkchar"; | ||
| 92 | + private String bpeVocab = ""; | ||
| 79 | 93 | ||
| 80 | public OnlineModelConfig build() { | 94 | public OnlineModelConfig build() { |
| 81 | return new OnlineModelConfig(this); | 95 | return new OnlineModelConfig(this); |
| @@ -125,5 +139,13 @@ public class OnlineModelConfig { | @@ -125,5 +139,13 @@ public class OnlineModelConfig { | ||
| 125 | this.modelType = modelType; | 139 | this.modelType = modelType; |
| 126 | return this; | 140 | return this; |
| 127 | } | 141 | } |
| 142 | + | ||
| 143 | + public void setModelingUnit(String modelingUnit) { | ||
| 144 | + this.modelingUnit = modelingUnit; | ||
| 145 | + } | ||
| 146 | + | ||
| 147 | + public void setBpeVocab(String bpeVocab) { | ||
| 148 | + this.bpeVocab = bpeVocab; | ||
| 149 | + } | ||
| 128 | } | 150 | } |
| 129 | } | 151 | } |
-
请 注册 或 登录 后发表评论