app.py

from huggingface_hub import from_pretrained_fastai
import gradio as gr
from fastai.text.all import *

# repo_id = "YOUR_USERNAME/YOUR_LEARNER_NAME"
repo_id = "nereamart/rotten_tomatoes"

# Cargar el modelo
learner = from_pretrained_fastai(repo_id)
labels = learner.dls.vocab

# Función para hacer predicciones de texto
def predict(text):
    pred = learner.predict(text)
    if pred == 1:
        return {"POSITIVE": 1.0, "NEGATIVE": 0.0}
    else:
        return {"POSITIVE": 0.0, "NEGATIVE": 1.0}

# Creamos la interfaz con componentes de texto
examples = [
    "This movie was fantastic! The acting was superb.",
    "Terrible film. Waste of time and money.",
    "It was okay, not great but not awful either."
]

interface = gr.Interface(
    fn=predict, 
    inputs=gr.Textbox(lines=5, placeholder="Review"), 
    outputs=gr.Label(num_top_classes=2),
    examples=examples,
    title="Rotten Tomatoes Review Classifier",
    description="Predict whether a movie review is positive (1) or negative (0)"
)

interface.launch(share=False)