Update app.py
Browse files
app.py
CHANGED
@@ -296,18 +296,40 @@ if st.button('Сгенерировать потери'):
|
|
296 |
|
297 |
r = speech_r.Recognizer()
|
298 |
|
|
|
|
|
|
|
299 |
harvard = speech_r.AudioFile('target.wav')
|
300 |
with harvard as source:
|
301 |
audio = r.record(source)
|
302 |
|
303 |
orig = r.recognize_google(audio, language = "ru-RU")
|
304 |
|
|
|
|
|
|
|
|
|
|
|
305 |
harvard = speech_r.AudioFile('lossy.wav')
|
306 |
with harvard as source:
|
307 |
audio = r.record(source)
|
308 |
lossy = r.recognize_google(audio, language = "ru-RU")
|
309 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
310 |
|
|
|
|
|
|
|
311 |
harvard = speech_r.AudioFile('enhanced.wav')
|
312 |
with harvard as source:
|
313 |
audio = r.record(source)
|
@@ -316,14 +338,14 @@ if st.button('Сгенерировать потери'):
|
|
316 |
error1 = wer(orig, orig)
|
317 |
error2 = wer(orig, lossy)
|
318 |
error3 = wer(orig, enhanced)
|
319 |
-
|
320 |
|
321 |
print(orig)
|
322 |
print(lossy)
|
323 |
print(enhanced)
|
324 |
st.text(orig)
|
325 |
st.text(lossy)
|
326 |
-
|
327 |
|
328 |
st.dataframe(df_1)
|
329 |
|
|
|
296 |
|
297 |
r = speech_r.Recognizer()
|
298 |
|
299 |
+
|
300 |
+
|
301 |
+
|
302 |
harvard = speech_r.AudioFile('target.wav')
|
303 |
with harvard as source:
|
304 |
audio = r.record(source)
|
305 |
|
306 |
orig = r.recognize_google(audio, language = "ru-RU")
|
307 |
|
308 |
+
|
309 |
+
|
310 |
+
|
311 |
+
|
312 |
+
|
313 |
harvard = speech_r.AudioFile('lossy.wav')
|
314 |
with harvard as source:
|
315 |
audio = r.record(source)
|
316 |
lossy = r.recognize_google(audio, language = "ru-RU")
|
317 |
|
318 |
+
try:
|
319 |
+
with harvard as source:
|
320 |
+
audio = r.record(source)
|
321 |
+
lossy = recognizer.recognize_google(audio)
|
322 |
+
#print("Распознанный текст:", text)
|
323 |
+
except sr.UnknownValueError:
|
324 |
+
print("Система не смогла распознать аудио")
|
325 |
+
lossy = ''
|
326 |
+
except sr.RequestError as e:
|
327 |
+
print("Ошибка при запросе к сервису распознавания речи; {0}".format(e))
|
328 |
+
|
329 |
|
330 |
+
|
331 |
+
|
332 |
+
|
333 |
harvard = speech_r.AudioFile('enhanced.wav')
|
334 |
with harvard as source:
|
335 |
audio = r.record(source)
|
|
|
338 |
error1 = wer(orig, orig)
|
339 |
error2 = wer(orig, lossy)
|
340 |
error3 = wer(orig, enhanced)
|
341 |
+
WER_mass=[error1*100, error2*100, error3*100]
|
342 |
|
343 |
print(orig)
|
344 |
print(lossy)
|
345 |
print(enhanced)
|
346 |
st.text(orig)
|
347 |
st.text(lossy)
|
348 |
+
df_1['WER'] = WER_mass
|
349 |
|
350 |
st.dataframe(df_1)
|
351 |
|