Update app.py

app.py

@@ -7,13 +7,11 @@ import gradio as gr
 import seaborn as sns
 import pandas as pd
 import plotly.express as px
-model = keras.models.load_model('/kaggle/input/emotions/emotion (1).h5')
+model = keras.models.load_model('emotion.h5')
 labels = ['Angry', 'Disgusted', 'Fearful', 'Happy', 'Neutral', 'Sad', 'Suprised']
 # def load()
 
 def predict(audio):
-#     audio_dir = ('/kaggle/input/sample-audios/')
-#     audio = os.path.join(audio_dir,audio_link)
     wave, sr = librosa.load(audio, sr=None)
     segment_dur_secs = 3  
     segment_length = sr * segment_dur_secs
@@ -42,8 +40,6 @@ def predict(audio):
                                       ,offset=0.5
                                      )
         sample_rate = np.array(sample_rate)
-
-        # mean as the feature. Could do min and max etc as well. 
         mfccs = np.mean(librosa.feature.mfcc(y=X, 
                                             sr=sample_rate, 
                                             n_mfcc=13),
@@ -64,11 +60,9 @@ def predict(audio):
     df_preds = df_preds.reset_index()
     import plotly.io as pio
     fig = px.line(df_preds, x="index", y="emotion", title='Life expectancy in Canada')
-#     plt = sns.lineplot(df_preds,x='index',y='emotion');
-#     plt.set_xlabel('samples(each in 3s interval)');
     return fig
 
-# outputs = gr.Plot()
+outputs = gr.Plot()
 title = "Emotion recognition"
 description = "This model can shows how speaker emotion changes over the speech"