tri_modelos / app.py
LRascon's picture
Creaci贸n de app
b8e69b5
raw
history blame
1.74 kB
import gradio as gr
import tensorflow as tf
from transformers import pipeline
inception_net = tf.keras.applications.MobileNetV2()
def clasificador_imagenes(inp):
inp = inp.reshape((-1, 224, 224, 3))
inp = tf.keras.applications.mobilenet_v2.preprocess_input(inp)
prediction = inception_net.predict(inp).reshape(1,1000)
pred_scores = tf.keras.applications.mobilenet_v2.decode_predictions(prediction, top=100)
confidence = {f'{pred_scores[0][i][1]}': float(pred_scores[0][i][2]) for i in range(100)}
return confidence
def audio_a_texto(audio):
text = trans(audio)["text"]
return text
def texto_a_sentimiento(text):
return classificator(text)[0]['label']
trans = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-large-xlsr-53-spanish")
classificator = pipeline("text-classification", model="pysentimiento/robertuito-sentiment-analysis")
demo = gr.Blocks()
with demo:
gr.Markdown("# Demo con Blocks")
with gr.Tabs():
with gr.TabItem("Transcribe Audio en espa帽ol"):
with gr.Row():
audio = gr.Audio(source='microphone', type='filepath')
transcript = gr.Textbox()
b1 = gr.Button("Transcribe")
with gr.TabItem("Analisis de sentimiento"):
with gr.Row():
texto = gr.Textbox()
label = gr.Label()
b2 = gr.Button("Sentimiento")
b1.click(audio_a_texto, inputs=audio, outputs=transcript)
b2.click(texto_a_sentimiento, inputs=texto, outputs=label)
with gr.TabItem("Clasificador de imagenes"):
with gr.Row():
image = gr.Image(shape=(224, 224))
label= gr.Label(num_top_classes=3)
bimage= gr.Button("Clasifica")
bimage.click(clasificador_imagenes, inputs=image, outputs=label)
demo.launch()