Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -91,8 +91,8 @@ def detect_onnx_models(path):
|
|
91 |
else:
|
92 |
return None
|
93 |
|
94 |
-
|
95 |
-
|
96 |
async def main(
|
97 |
request: Request,
|
98 |
text_input: str = Form(...),
|
@@ -189,10 +189,9 @@ async def main(
|
|
189 |
inferencing(model, config, sid, text, rate, noise_scale, noise_scale_w, auto_play)
|
190 |
|
191 |
# Save the audio as a temporary WAV file
|
192 |
-
|
193 |
-
|
194 |
-
|
195 |
-
# return {"message": f"Text to synthesize: {text_input}, Speed: {speed_slider}, Play: {play}"}
|
196 |
|
197 |
def load_onnx(model, sess_options, providers = ["CPUExecutionProvider"]):
|
198 |
_LOGGER.debug("Loading model from %s", model)
|
@@ -242,6 +241,7 @@ def phonemes_to_ids(config, phonemes: List[str]) -> List[int]:
|
|
242 |
ids.extend(id_map[EOS])
|
243 |
return ids
|
244 |
|
|
|
245 |
def inferencing(model, config, sid, line, length_scale = 1, noise_scale = 0.667, noise_scale_w = 0.8, auto_play=True):
|
246 |
audios = []
|
247 |
if config["phoneme_type"] == "PhonemeType.ESPEAK":
|
@@ -276,7 +276,9 @@ def inferencing(model, config, sid, line, length_scale = 1, noise_scale = 0.667,
|
|
276 |
audios.append(audio)
|
277 |
merged_audio = np.concatenate(audios)
|
278 |
sample_rate = config["audio"]["sample_rate"]
|
279 |
-
|
|
|
|
|
280 |
# Return the audio file as a FastAPI response
|
281 |
# display(Markdown(f"{line}"))
|
282 |
# display(Audio(merged_audio, rate=sample_rate, autoplay=auto_play))
|
|
|
91 |
else:
|
92 |
return None
|
93 |
|
94 |
+
@app.post("/synthesize")
|
95 |
+
#@app.post("/synthesize", response_class=FileResponse)
|
96 |
async def main(
|
97 |
request: Request,
|
98 |
text_input: str = Form(...),
|
|
|
189 |
inferencing(model, config, sid, text, rate, noise_scale, noise_scale_w, auto_play)
|
190 |
|
191 |
# Save the audio as a temporary WAV file
|
192 |
+
return {"message": "Post request received", "data": data}
|
193 |
+
|
194 |
+
# return {"message": f"Text to synthesize: {text_input}, Speed: {speed_slider}, Play: {play}"}
|
|
|
195 |
|
196 |
def load_onnx(model, sess_options, providers = ["CPUExecutionProvider"]):
|
197 |
_LOGGER.debug("Loading model from %s", model)
|
|
|
241 |
ids.extend(id_map[EOS])
|
242 |
return ids
|
243 |
|
244 |
+
@app.get(generate_audio)
|
245 |
def inferencing(model, config, sid, line, length_scale = 1, noise_scale = 0.667, noise_scale_w = 0.8, auto_play=True):
|
246 |
audios = []
|
247 |
if config["phoneme_type"] == "PhonemeType.ESPEAK":
|
|
|
276 |
audios.append(audio)
|
277 |
merged_audio = np.concatenate(audios)
|
278 |
sample_rate = config["audio"]["sample_rate"]
|
279 |
+
temp_audio_path = os.path.join(tempfile.gettempdir(), "generated_audio.wav")
|
280 |
+
sf.write(temp_audio_path, merged_audio, config["audio"]["sample_rate"])
|
281 |
+
return FileResponse(temp_audio_path)
|
282 |
# Return the audio file as a FastAPI response
|
283 |
# display(Markdown(f"{line}"))
|
284 |
# display(Audio(merged_audio, rate=sample_rate, autoplay=auto_play))
|