Spaces:
Running
Running
csukuangfj
commited on
Commit
•
abeb4dc
1
Parent(s):
71c0d81
add melo tts
Browse files- model.py +10 -1
- requirements.txt +2 -2
model.py
CHANGED
@@ -236,6 +236,8 @@ def _get_vits_hf(repo_id: str, speed: float) -> sherpa_onnx.OfflineTts:
|
|
236 |
|
237 |
if "fanchen" in repo_id or "vits-cantonese-hf-xiaomaiiwn" in repo_id:
|
238 |
model = repo_id.split("/")[-1]
|
|
|
|
|
239 |
else:
|
240 |
model = repo_id.split("-")[-1]
|
241 |
|
@@ -269,7 +271,7 @@ def _get_vits_hf(repo_id: str, speed: float) -> sherpa_onnx.OfflineTts:
|
|
269 |
rule_fars = ""
|
270 |
|
271 |
if "vits-cantonese-hf-xiaomaiiwn" not in repo_id:
|
272 |
-
rule_fsts = ["phone.fst", "date.fst", "number.fst"
|
273 |
|
274 |
rule_fsts = [
|
275 |
get_file(
|
@@ -320,6 +322,8 @@ def _get_vits_hf(repo_id: str, speed: float) -> sherpa_onnx.OfflineTts:
|
|
320 |
def get_pretrained_model(repo_id: str, speed: float) -> sherpa_onnx.OfflineTts:
|
321 |
if repo_id in chinese_models:
|
322 |
return chinese_models[repo_id](repo_id, speed)
|
|
|
|
|
323 |
if repo_id in cantonese_models:
|
324 |
return cantonese_models[repo_id](repo_id, speed)
|
325 |
elif repo_id in english_models:
|
@@ -424,6 +428,10 @@ cantonese_models = {
|
|
424 |
"csukuangfj/vits-cantonese-hf-xiaomaiiwn": _get_vits_hf,
|
425 |
}
|
426 |
|
|
|
|
|
|
|
|
|
427 |
chinese_models = {
|
428 |
"csukuangfj/vits-zh-hf-fanchen-wnj|1": _get_vits_hf, # 1
|
429 |
"csukuangfj/vits-zh-hf-fanchen-C|187": _get_vits_hf, # 187
|
@@ -766,6 +774,7 @@ welsh_models = {
|
|
766 |
language_to_models = {
|
767 |
"English": list(english_models.keys()),
|
768 |
"Chinese (Mandarin, 普通话)": list(chinese_models.keys()),
|
|
|
769 |
"Cantonese (粤语)": list(cantonese_models.keys()),
|
770 |
"Min-nan (闽南话)": list(min_nan_models.keys()),
|
771 |
"Arabic": list(arabic_models.keys()),
|
|
|
236 |
|
237 |
if "fanchen" in repo_id or "vits-cantonese-hf-xiaomaiiwn" in repo_id:
|
238 |
model = repo_id.split("/")[-1]
|
239 |
+
elif "csukuangfj/vits-melo-tts-zh_en" == repo_id:
|
240 |
+
model = "model"
|
241 |
else:
|
242 |
model = repo_id.split("-")[-1]
|
243 |
|
|
|
271 |
rule_fars = ""
|
272 |
|
273 |
if "vits-cantonese-hf-xiaomaiiwn" not in repo_id:
|
274 |
+
rule_fsts = ["phone.fst", "date.fst", "number.fst"]
|
275 |
|
276 |
rule_fsts = [
|
277 |
get_file(
|
|
|
322 |
def get_pretrained_model(repo_id: str, speed: float) -> sherpa_onnx.OfflineTts:
|
323 |
if repo_id in chinese_models:
|
324 |
return chinese_models[repo_id](repo_id, speed)
|
325 |
+
elif repo_id in chinese_english_models:
|
326 |
+
return chinese_english_models[repo_id](repo_id, speed)
|
327 |
if repo_id in cantonese_models:
|
328 |
return cantonese_models[repo_id](repo_id, speed)
|
329 |
elif repo_id in english_models:
|
|
|
428 |
"csukuangfj/vits-cantonese-hf-xiaomaiiwn": _get_vits_hf,
|
429 |
}
|
430 |
|
431 |
+
chinese_english_models = {
|
432 |
+
"csukuangfj/vits-melo-tts-zh_en|1": _get_vits_hf, # 1
|
433 |
+
}
|
434 |
+
|
435 |
chinese_models = {
|
436 |
"csukuangfj/vits-zh-hf-fanchen-wnj|1": _get_vits_hf, # 1
|
437 |
"csukuangfj/vits-zh-hf-fanchen-C|187": _get_vits_hf, # 187
|
|
|
774 |
language_to_models = {
|
775 |
"English": list(english_models.keys()),
|
776 |
"Chinese (Mandarin, 普通话)": list(chinese_models.keys()),
|
777 |
+
"Chinese+English": list(chinese_english_models.keys()),
|
778 |
"Cantonese (粤语)": list(cantonese_models.keys()),
|
779 |
"Min-nan (闽南话)": list(min_nan_models.keys()),
|
780 |
"Arabic": list(arabic_models.keys()),
|
requirements.txt
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
https://huggingface.co/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-1.9.23-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
|
2 |
-
|
3 |
|
4 |
soundfile
|
|
|
1 |
+
#https://huggingface.co/csukuangfj/sherpa-onnx-wheels/resolve/main/sherpa_onnx-1.9.23-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
|
2 |
+
sherpa-onnx>=1.10.16
|
3 |
|
4 |
soundfile
|