use filepath
Browse files
app.py
CHANGED
@@ -59,12 +59,25 @@ def tts(text, speaker_idx):
|
|
59 |
wav_coll = model_collectivat.tts(text)
|
60 |
wav_piper = model_piper.synthesize(text)
|
61 |
|
62 |
-
return (model_bsc.tts_config.audio["sample_rate"], wav_bsc), (22000, wav_coll), (16000, wav_piper)
|
63 |
|
64 |
# return output
|
|
|
65 |
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
|
66 |
-
model_bsc.save_wav(
|
67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
|
69 |
|
70 |
description="""
|
@@ -83,12 +96,12 @@ iface = gr.Interface(
|
|
83 |
label="Text",
|
84 |
default="L'Èlia i l'Alí a l'aula. L'oli i l'ou. Lulú olorava la lila.",
|
85 |
),
|
86 |
-
gr.Dropdown(label="Selecciona un parlant", choices=SPEAKERS, default=
|
87 |
],
|
88 |
outputs=[
|
89 |
-
gr.Audio(label="BSC VITS",type="
|
90 |
-
gr.Audio(label="Collectivat Fastspeech",type="
|
91 |
-
gr.Audio(label="Piper VITS",type="
|
92 |
],
|
93 |
title="🗣️ TTS Català Multi Parlant - VITS 🗣️",
|
94 |
description=description,
|
|
|
59 |
wav_coll = model_collectivat.tts(text)
|
60 |
wav_piper = model_piper.synthesize(text)
|
61 |
|
62 |
+
#return (model_bsc.tts_config.audio["sample_rate"], wav_bsc), (22000, wav_coll), (16000, wav_piper)
|
63 |
|
64 |
# return output
|
65 |
+
fp_bsc = ""
|
66 |
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
|
67 |
+
model_bsc.save_wav(wav_bsc, fp)
|
68 |
+
fp_bsc = fp.name
|
69 |
+
|
70 |
+
fp_coll = ""
|
71 |
+
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
|
72 |
+
model_collectivat.save_wav(wav_coll, fp)
|
73 |
+
fp_coll = fp.name
|
74 |
+
|
75 |
+
fp_piper = ""
|
76 |
+
with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
|
77 |
+
fp.write(wav_piper)
|
78 |
+
fp_piper = fp.name
|
79 |
+
|
80 |
+
return fp_bsc, fp_coll, fp_piper
|
81 |
|
82 |
|
83 |
description="""
|
|
|
96 |
label="Text",
|
97 |
default="L'Èlia i l'Alí a l'aula. L'oli i l'ou. Lulú olorava la lila.",
|
98 |
),
|
99 |
+
gr.Dropdown(label="Selecciona un parlant", choices=SPEAKERS, default="ona")
|
100 |
],
|
101 |
outputs=[
|
102 |
+
gr.Audio(label="BSC VITS",type="filepath"),
|
103 |
+
gr.Audio(label="Collectivat Fastspeech",type="filepath"),
|
104 |
+
gr.Audio(label="Piper VITS",type="filepath")
|
105 |
],
|
106 |
title="🗣️ TTS Català Multi Parlant - VITS 🗣️",
|
107 |
description=description,
|
engine.py
CHANGED
@@ -104,7 +104,7 @@ class Piper:
|
|
104 |
},
|
105 |
)[0].squeeze((0, 1))
|
106 |
audio = audio_float_to_int16(audio.squeeze())
|
107 |
-
|
108 |
# Convert to WAV
|
109 |
with io.BytesIO() as wav_io:
|
110 |
wav_file: wave.Wave_write = wave.open(wav_io, "wb")
|
|
|
104 |
},
|
105 |
)[0].squeeze((0, 1))
|
106 |
audio = audio_float_to_int16(audio.squeeze())
|
107 |
+
|
108 |
# Convert to WAV
|
109 |
with io.BytesIO() as wav_io:
|
110 |
wav_file: wave.Wave_write = wave.open(wav_io, "wb")
|