aadarsh_seamless_m4t / lang_list.py
aadarsh-lalchandani's picture
Update lang_list.py
a941097
raw
history blame
984 Bytes
# Language dict
language_code_to_name = {
"ben": "Bengali",
"eng": "English",
"guj": "Gujarati",
"hin": "Hindi",
"snd": "Sindhi",
}
LANGUAGE_NAME_TO_CODE = {v: k for k, v in language_code_to_name.items()}
# Source langs: S2ST / S2TT / ASR don't need source lang
# T2TT / T2ST use this
text_source_language_codes = [
"ben",
"eng",
"guj",
"hin",
"snd",
]
TEXT_SOURCE_LANGUAGE_NAMES = sorted([language_code_to_name[code] for code in text_source_language_codes])
# Target langs:
# S2ST / T2ST
s2st_target_language_codes = [
"ben",
"eng",
"guj",
"hin",
"snd",
]
S2ST_TARGET_LANGUAGE_NAMES = sorted([language_code_to_name[code] for code in s2st_target_language_codes])
# S2TT / ASR
S2TT_TARGET_LANGUAGE_NAMES = TEXT_SOURCE_LANGUAGE_NAMES
# T2TT
T2TT_TARGET_LANGUAGE_NAMES = TEXT_SOURCE_LANGUAGE_NAMES
LANG_TO_SPKR_ID = {
"ben": [
2,
1
],
"eng": [
10
],
"hin": [
16
],
}