Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -20,7 +20,7 @@ pipeline = PIPELINE(model, "20B_tokenizer.json")
|
|
20 |
from TTS.api import TTS
|
21 |
tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=True)
|
22 |
import whisper
|
23 |
-
|
24 |
|
25 |
os.system('pip install voicefixer --upgrade')
|
26 |
from voicefixer import VoiceFixer
|
@@ -66,16 +66,16 @@ def evaluate(
|
|
66 |
audio = whisper.load_audio(audio)
|
67 |
audio = whisper.pad_or_trim(audio)
|
68 |
|
69 |
-
# make log-Mel spectrogram and move to the same device as the
|
70 |
-
mel = whisper.log_mel_spectrogram(audio).to(
|
71 |
|
72 |
# detect the spoken language
|
73 |
-
_, probs =
|
74 |
print(f"Detected language: {max(probs, key=probs.get)}")
|
75 |
|
76 |
# decode the audio
|
77 |
options = whisper.DecodingOptions()
|
78 |
-
result = whisper.decode(
|
79 |
|
80 |
|
81 |
res = []
|
@@ -122,9 +122,9 @@ def evaluate(
|
|
122 |
|
123 |
res.append(out_str.strip())
|
124 |
|
125 |
-
res1 = ''.join(str(x) for x in res)
|
126 |
|
127 |
-
tts.tts_to_file(
|
128 |
|
129 |
voicefixer.restore(input="output.wav", # input wav file path
|
130 |
output="audio1.wav", # output wav file path
|
|
|
20 |
from TTS.api import TTS
|
21 |
tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=True)
|
22 |
import whisper
|
23 |
+
model1 = whisper.load_model("small")
|
24 |
|
25 |
os.system('pip install voicefixer --upgrade')
|
26 |
from voicefixer import VoiceFixer
|
|
|
66 |
audio = whisper.load_audio(audio)
|
67 |
audio = whisper.pad_or_trim(audio)
|
68 |
|
69 |
+
# make log-Mel spectrogram and move to the same device as the model1
|
70 |
+
mel = whisper.log_mel_spectrogram(audio).to(model1.device)
|
71 |
|
72 |
# detect the spoken language
|
73 |
+
_, probs = model1.detect_language(mel)
|
74 |
print(f"Detected language: {max(probs, key=probs.get)}")
|
75 |
|
76 |
# decode the audio
|
77 |
options = whisper.DecodingOptions()
|
78 |
+
result = whisper.decode(model1, mel, options)
|
79 |
|
80 |
|
81 |
res = []
|
|
|
122 |
|
123 |
res.append(out_str.strip())
|
124 |
|
125 |
+
# res1 = ''.join(str(x) for x in res)
|
126 |
|
127 |
+
tts.tts_to_file(res, speaker_wav = upload, language="en", file_path="output.wav")
|
128 |
|
129 |
voicefixer.restore(input="output.wav", # input wav file path
|
130 |
output="audio1.wav", # output wav file path
|