holylovenia
commited on
Upload jv_id_tts.py with huggingface_hub
Browse files- jv_id_tts.py +15 -15
jv_id_tts.py
CHANGED
@@ -5,14 +5,14 @@ from typing import List
|
|
5 |
|
6 |
import datasets
|
7 |
|
8 |
-
from
|
9 |
-
from
|
10 |
-
from
|
11 |
DEFAULT_SOURCE_VIEW_NAME, Tasks)
|
12 |
|
13 |
_DATASETNAME = "jv_id_tts"
|
14 |
_SOURCE_VIEW_NAME = DEFAULT_SOURCE_VIEW_NAME
|
15 |
-
_UNIFIED_VIEW_NAME =
|
16 |
|
17 |
_LANGUAGES = ["jav"]
|
18 |
_LOCAL = False
|
@@ -38,7 +38,7 @@ This dataset was collected by Google in collaboration with Gadjah Mada Universit
|
|
38 |
|
39 |
_HOMEPAGE = "http://openslr.org/41/"
|
40 |
|
41 |
-
_LICENSE =
|
42 |
|
43 |
_URLs = {
|
44 |
_DATASETNAME: {
|
@@ -50,25 +50,25 @@ _URLs = {
|
|
50 |
_SUPPORTED_TASKS = [Tasks.TEXT_TO_SPEECH]
|
51 |
|
52 |
_SOURCE_VERSION = "1.0.0"
|
53 |
-
|
54 |
|
55 |
|
56 |
class JvIdTTS(datasets.GeneratorBasedBuilder):
|
57 |
"""jv_id_tts contains high-quality Multi-speaker TTS data for Javanese (jv-ID)."""
|
58 |
|
59 |
BUILDER_CONFIGS = [
|
60 |
-
|
61 |
name="jv_id_tts_source",
|
62 |
version=datasets.Version(_SOURCE_VERSION),
|
63 |
description="JV_ID_TTS source schema",
|
64 |
schema="source",
|
65 |
subset_id="jv_id_tts",
|
66 |
),
|
67 |
-
|
68 |
-
name="
|
69 |
-
version=datasets.Version(
|
70 |
description="JV_ID_TTS Nusantara schema",
|
71 |
-
schema="
|
72 |
subset_id="jv_id_tts",
|
73 |
),
|
74 |
]
|
@@ -86,7 +86,7 @@ class JvIdTTS(datasets.GeneratorBasedBuilder):
|
|
86 |
"text": datasets.Value("string"),
|
87 |
}
|
88 |
)
|
89 |
-
elif self.config.schema == "
|
90 |
features = schemas.speech_text_features
|
91 |
|
92 |
return datasets.DatasetInfo(
|
@@ -114,7 +114,7 @@ class JvIdTTS(datasets.GeneratorBasedBuilder):
|
|
114 |
|
115 |
def _generate_examples(self, male_filepath: Path, female_filepath: Path):
|
116 |
|
117 |
-
if self.config.schema == "source" or self.config.schema == "
|
118 |
tsv_file = os.path.join(male_filepath, "jv_id_male", "line_index.tsv")
|
119 |
with open(tsv_file, "r") as file:
|
120 |
tsv_data = csv.reader(file, delimiter="\t")
|
@@ -135,7 +135,7 @@ class JvIdTTS(datasets.GeneratorBasedBuilder):
|
|
135 |
"text": transcription_text,
|
136 |
}
|
137 |
yield audio_id, ex
|
138 |
-
elif self.config.schema == "
|
139 |
ex = {
|
140 |
"id": audio_id,
|
141 |
"speaker_id": speaker_id,
|
@@ -168,7 +168,7 @@ class JvIdTTS(datasets.GeneratorBasedBuilder):
|
|
168 |
"text": transcription_text,
|
169 |
}
|
170 |
yield audio_id, ex
|
171 |
-
elif self.config.schema == "
|
172 |
ex = {
|
173 |
"id": audio_id,
|
174 |
"speaker_id": speaker_id,
|
|
|
5 |
|
6 |
import datasets
|
7 |
|
8 |
+
from seacrowd.utils import schemas
|
9 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
10 |
+
from seacrowd.utils.constants import (DEFAULT_SEACROWD_VIEW_NAME, Licenses,
|
11 |
DEFAULT_SOURCE_VIEW_NAME, Tasks)
|
12 |
|
13 |
_DATASETNAME = "jv_id_tts"
|
14 |
_SOURCE_VIEW_NAME = DEFAULT_SOURCE_VIEW_NAME
|
15 |
+
_UNIFIED_VIEW_NAME = DEFAULT_SEACROWD_VIEW_NAME
|
16 |
|
17 |
_LANGUAGES = ["jav"]
|
18 |
_LOCAL = False
|
|
|
38 |
|
39 |
_HOMEPAGE = "http://openslr.org/41/"
|
40 |
|
41 |
+
_LICENSE = Licenses.CC_BY_SA_4_0.value
|
42 |
|
43 |
_URLs = {
|
44 |
_DATASETNAME: {
|
|
|
50 |
_SUPPORTED_TASKS = [Tasks.TEXT_TO_SPEECH]
|
51 |
|
52 |
_SOURCE_VERSION = "1.0.0"
|
53 |
+
_SEACROWD_VERSION = "2024.06.20"
|
54 |
|
55 |
|
56 |
class JvIdTTS(datasets.GeneratorBasedBuilder):
|
57 |
"""jv_id_tts contains high-quality Multi-speaker TTS data for Javanese (jv-ID)."""
|
58 |
|
59 |
BUILDER_CONFIGS = [
|
60 |
+
SEACrowdConfig(
|
61 |
name="jv_id_tts_source",
|
62 |
version=datasets.Version(_SOURCE_VERSION),
|
63 |
description="JV_ID_TTS source schema",
|
64 |
schema="source",
|
65 |
subset_id="jv_id_tts",
|
66 |
),
|
67 |
+
SEACrowdConfig(
|
68 |
+
name="jv_id_tts_seacrowd_sptext",
|
69 |
+
version=datasets.Version(_SEACROWD_VERSION),
|
70 |
description="JV_ID_TTS Nusantara schema",
|
71 |
+
schema="seacrowd_sptext",
|
72 |
subset_id="jv_id_tts",
|
73 |
),
|
74 |
]
|
|
|
86 |
"text": datasets.Value("string"),
|
87 |
}
|
88 |
)
|
89 |
+
elif self.config.schema == "seacrowd_sptext":
|
90 |
features = schemas.speech_text_features
|
91 |
|
92 |
return datasets.DatasetInfo(
|
|
|
114 |
|
115 |
def _generate_examples(self, male_filepath: Path, female_filepath: Path):
|
116 |
|
117 |
+
if self.config.schema == "source" or self.config.schema == "seacrowd_sptext":
|
118 |
tsv_file = os.path.join(male_filepath, "jv_id_male", "line_index.tsv")
|
119 |
with open(tsv_file, "r") as file:
|
120 |
tsv_data = csv.reader(file, delimiter="\t")
|
|
|
135 |
"text": transcription_text,
|
136 |
}
|
137 |
yield audio_id, ex
|
138 |
+
elif self.config.schema == "seacrowd_sptext":
|
139 |
ex = {
|
140 |
"id": audio_id,
|
141 |
"speaker_id": speaker_id,
|
|
|
168 |
"text": transcription_text,
|
169 |
}
|
170 |
yield audio_id, ex
|
171 |
+
elif self.config.schema == "seacrowd_sptext":
|
172 |
ex = {
|
173 |
"id": audio_id,
|
174 |
"speaker_id": speaker_id,
|