Update app.py

app.py

@@ -22,18 +22,9 @@ def predict_and_ctc_lm_decode(input_file):
     speech = load_and_fix_data(input_file, sampling_rate)
     transcribed_text = asr(speech, chunk_length_s=5, stride_length_s=1)["text"]
     pipe1 = pipeline("sentiment-analysis", model = "finiteautomata/beto-sentiment-analysis")
-    sentiment = pipe1(transcribed_text)
-    sentiment={dic["label"]: dic["score"] for dic in sentiment}
-    pipe2 = pipeline("text-classification", model = "hackathon-pln-es/twitter_sexismo-finetuned-robertuito-exist2021")
-    sexism_detection = pipe2(transcribed_text)
-    sexism_detection={dic["label"]: dic["score"] for dic in sexism_detection}
-    #sexism_detection = np.where(sexism_detection['label']== 0, 'No Sexista', 'Sexista')
-    pipe3 = pipeline("text-classification", model = "hackathon-pln-es/twitter_sexismo-finetuned-robertuito-exist2021")
-    harassment_detection = pipe3(transcribed_text)
-    harassment_detection={dic["label"]: dic["score"] for dic in harassment_detection}
-    #harassment_detection = np.where(harassment_detection['label']== 0, 'No Harassment', 'Harassment')
+    sentiment = pipe1(transcribed_text)[0]["label"]
     return sentiment
-#sexism_detection, harassment_detection
+
 
 gr.Interface(
     predict_and_ctc_lm_decode,
@@ -41,7 +32,7 @@ gr.Interface(
         gr.inputs.Audio(source="microphone", type="filepath", label="Record your audio")
     ],
     #outputs=[gr.outputs.Label(num_top_classes=2),gr.outputs.Label(num_top_classes=2), gr.outputs.Label(num_top_classes=2)],
-    outputs=[gr.outputs.Label(num_top_classes=2)],
+    outputs=[gr.outputs.Textbox(label="Predicción")],
     examples=[["audio_test.wav"], ["sample_audio.wav"]],
     title="Sentiment Analysis of Spanish Transcribed Audio",
     description="This is a Gradio demo for Sentiment Analysis of Transcribed Spanish Audio. First, we do Speech to Text, and then we perform sentiment analysis on the obtained transcription of the input audio.",