Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -135,6 +135,42 @@ def percent_V(vowels, total_wo_pauses):
|
|
135 |
return pV
|
136 |
|
137 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
138 |
|
139 |
def transcribe(audio):
|
140 |
y, sr = sf.read(audio)
|
@@ -163,7 +199,7 @@ def transcribe(audio):
|
|
163 |
dC = delta_C(cons_clusters)
|
164 |
pV = percent_V(vowels_duration, duration_without_pauses)
|
165 |
|
166 |
-
transcription = processor.decode(predicted_ids
|
167 |
|
168 |
text = {"transcription": transcription}
|
169 |
|
@@ -178,7 +214,7 @@ iface = gr.Interface(
|
|
178 |
inputs=gr.Audio(type="filepath"),
|
179 |
outputs="text",
|
180 |
title="Mihaj/Wav2Vec2RhytmAnalyzer",
|
181 |
-
description="Демо анализатор ритма на основе модели Wav2Vec large от bond005.
|
182 |
)
|
183 |
|
184 |
iface.launch()
|
|
|
135 |
return pV
|
136 |
|
137 |
|
138 |
+
# point_1 = np.array((0, 0, 0))
|
139 |
+
# point_2 = np.array((3, 3, 3))
|
140 |
+
def count_eucl(point_1, point_2):
|
141 |
+
# Initializing the points
|
142 |
+
# Get the square of the difference of the 2 vectors
|
143 |
+
square = np.square(point_1 - point_2)
|
144 |
+
# Get the sum of the square
|
145 |
+
sum_square = np.sum(square)
|
146 |
+
# The last step is to get the square root and print the Euclidean distance
|
147 |
+
distance = np.sqrt(sum_square)
|
148 |
+
return distance
|
149 |
+
|
150 |
+
|
151 |
+
ex_dict = {"eng": np.array((0.0535, 0.401)), "kat": np.array((0.0452, 0.456)), "jap": np.array((0.0356, 0.531))}
|
152 |
+
|
153 |
+
|
154 |
+
def classify_rhytm(dC, pV):
|
155 |
+
our = np.array((dC, pV))
|
156 |
+
res = {}
|
157 |
+
if (dC > 0.08 and pV > 0.45) or (dC < 0.03 and pV < 0.04):
|
158 |
+
text = "Вы не укладываетесь ни в какие рамки и прекрасны в этом!"
|
159 |
+
else:
|
160 |
+
for k, v in ex_dict.items():
|
161 |
+
res[k] = count_eucl(our, v)
|
162 |
+
|
163 |
+
sorted_tuples = sorted(res.items(), key=lambda item: item[1])
|
164 |
+
sorted_res = {k: v for k, v in sorted_tuples}
|
165 |
+
if sorted_res.keys()[0] == "eng":
|
166 |
+
text = "По типу ритма ваша речь близка к тактосчитающим языкам (английский)."
|
167 |
+
if sorted_res.keys()[0] == "kat":
|
168 |
+
text = "По типу ритма ваша речь близка к слогосчитающим языкам (испанский)."
|
169 |
+
if sorted_res.keys()[0] == "jap":
|
170 |
+
text = "По типу ритма ваша речь близка к моросчитающим языкам (японский)."
|
171 |
+
return text
|
172 |
+
|
173 |
+
|
174 |
|
175 |
def transcribe(audio):
|
176 |
y, sr = sf.read(audio)
|
|
|
199 |
dC = delta_C(cons_clusters)
|
200 |
pV = percent_V(vowels_duration, duration_without_pauses)
|
201 |
|
202 |
+
transcription = processor.decode(predicted_ids).lower()
|
203 |
|
204 |
text = {"transcription": transcription}
|
205 |
|
|
|
214 |
inputs=gr.Audio(type="filepath"),
|
215 |
outputs="text",
|
216 |
title="Mihaj/Wav2Vec2RhytmAnalyzer",
|
217 |
+
description="Демо анализатор ритма на основе модели Wav2Vec large от bond005.",
|
218 |
)
|
219 |
|
220 |
iface.launch()
|