Update app.py
Browse files
app.py
CHANGED
@@ -1,50 +1,57 @@
|
|
1 |
import gradio as gr
|
2 |
import torchaudio
|
3 |
from transformers import pipeline
|
4 |
-
from datasets import load_dataset, Audio
|
5 |
|
6 |
# Load your model
|
7 |
classifier = pipeline("audio-classification", model="Ahmed107/Hamsa-Conversational-v1.0-mulaw-eos-v3-mulaw")
|
8 |
|
9 |
-
# Function to resample audio to 16kHz
|
10 |
def resample_audio(audio_file, target_sampling_rate=16000):
|
11 |
waveform, original_sample_rate = torchaudio.load(audio_file)
|
|
|
|
|
12 |
if original_sample_rate != target_sampling_rate:
|
13 |
resampler = torchaudio.transforms.Resample(orig_freq=original_sample_rate, new_freq=target_sampling_rate)
|
14 |
waveform = resampler(waveform)
|
|
|
|
|
|
|
|
|
|
|
15 |
return waveform.squeeze().numpy(), target_sampling_rate
|
16 |
|
17 |
# Define the prediction function
|
18 |
def classify_audio(audio_file):
|
19 |
-
# Resample the audio to 16kHz
|
20 |
-
resampled_audio,
|
21 |
|
22 |
-
#
|
23 |
-
|
|
|
24 |
|
25 |
-
# Return predictions as a dictionary
|
26 |
return {entry['label']: entry['score'] for entry in prediction}
|
27 |
|
28 |
# Define Gradio interface
|
29 |
def demo():
|
30 |
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
31 |
-
gr.Markdown("## Eos")
|
32 |
|
33 |
-
# Input Audio
|
34 |
with gr.Row():
|
35 |
audio_input = gr.Audio(type="filepath", label="Input Audio")
|
36 |
|
37 |
-
# Output Labels
|
38 |
with gr.Row():
|
39 |
label_output = gr.Label(label="Prediction")
|
40 |
|
41 |
# Predict Button
|
42 |
classify_btn = gr.Button("Classify")
|
43 |
-
|
44 |
-
#
|
45 |
classify_btn.click(fn=classify_audio, inputs=audio_input, outputs=label_output)
|
46 |
|
47 |
return demo
|
48 |
|
49 |
-
# Launch the demo
|
50 |
-
demo().launch()
|
|
|
1 |
import gradio as gr
|
2 |
import torchaudio
|
3 |
from transformers import pipeline
|
|
|
4 |
|
5 |
# Load your model
|
6 |
classifier = pipeline("audio-classification", model="Ahmed107/Hamsa-Conversational-v1.0-mulaw-eos-v3-mulaw")
|
7 |
|
8 |
+
# Function to resample audio to 16kHz and convert to mono if needed
|
9 |
def resample_audio(audio_file, target_sampling_rate=16000):
|
10 |
waveform, original_sample_rate = torchaudio.load(audio_file)
|
11 |
+
|
12 |
+
# Resample if necessary
|
13 |
if original_sample_rate != target_sampling_rate:
|
14 |
resampler = torchaudio.transforms.Resample(orig_freq=original_sample_rate, new_freq=target_sampling_rate)
|
15 |
waveform = resampler(waveform)
|
16 |
+
|
17 |
+
# Convert stereo to mono by averaging channels (if needed)
|
18 |
+
if waveform.shape[0] > 1:
|
19 |
+
waveform = waveform.mean(dim=0, keepdim=True)
|
20 |
+
|
21 |
return waveform.squeeze().numpy(), target_sampling_rate
|
22 |
|
23 |
# Define the prediction function
|
24 |
def classify_audio(audio_file):
|
25 |
+
# Resample the audio to 16kHz and handle channels
|
26 |
+
resampled_audio, sampling_rate = resample_audio(audio_file)
|
27 |
|
28 |
+
# Pass both the array and sampling_rate to the classifier
|
29 |
+
input_audio = {"array": resampled_audio, "sampling_rate": sampling_rate}
|
30 |
+
prediction = classifier(input_audio)
|
31 |
|
32 |
+
# Return predictions as a dictionary with labels and their scores
|
33 |
return {entry['label']: entry['score'] for entry in prediction}
|
34 |
|
35 |
# Define Gradio interface
|
36 |
def demo():
|
37 |
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
38 |
+
gr.Markdown("## Eos Audio Classification")
|
39 |
|
40 |
+
# Input Audio component
|
41 |
with gr.Row():
|
42 |
audio_input = gr.Audio(type="filepath", label="Input Audio")
|
43 |
|
44 |
+
# Output Labels component
|
45 |
with gr.Row():
|
46 |
label_output = gr.Label(label="Prediction")
|
47 |
|
48 |
# Predict Button
|
49 |
classify_btn = gr.Button("Classify")
|
50 |
+
|
51 |
+
# Set the button click action
|
52 |
classify_btn.click(fn=classify_audio, inputs=audio_input, outputs=label_output)
|
53 |
|
54 |
return demo
|
55 |
|
56 |
+
# Launch the Gradio demo
|
57 |
+
demo().launch()
|