app.py

import torch
from transformers import RobertaTokenizer, RobertaForSequenceClassification, Trainer, TrainingArguments
from datasets import load_dataset, load_metric
import gradio as gr

# Carregar o dataset IMDB
dataset = load_dataset('imdb')
metric = load_metric('accuracy')

# Carregar o tokenizer e o modelo RoBERTa
tokenizer = RobertaTokenizer.from_pretrained('roberta-base')
model = RobertaForSequenceClassification.from_pretrained('roberta-base')

# Tokenizar os dados
def preprocess_function(examples):
    return tokenizer(examples['text'], padding='max_length', truncation=True)

tokenized_datasets = dataset.map(preprocess_function, batched=True)

# Preparar o data collator
from transformers import DataCollatorWithPadding
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)

# Configurar os argumentos de treinamento
training_args = TrainingArguments(
    output_dir='./results',
    evaluation_strategy='epoch',
    per_device_train_batch_size=8,
    per_device_eval_batch_size=8,
    num_train_epochs=3,
    weight_decay=0.01,
)

# Definir a função de métricas
def compute_metrics(eval_pred):
    logits, labels = eval_pred
    predictions = torch.argmax(logits, dim=-1)
    return metric.compute(predictions=predictions, references=labels)

# Definir o Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_datasets['train'],
    eval_dataset=tokenized_datasets['test'],
    tokenizer=tokenizer,
    data_collator=data_collator,
    compute_metrics=compute_metrics
)

# Treinar o modelo
trainer.train()

# Avaliar o modelo
results = trainer.evaluate()
print(results)

# Salvar o modelo
model.save_pretrained('./model')
tokenizer.save_pretrained('./model')

# Função de inferência
def predict(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
    outputs = model(**inputs)
    predictions = torch.argmax(outputs.logits, dim=-1)
    return "Positive" if predictions.item() == 1 else "Negative"

# Interface Gradio
iface = gr.Interface(
    fn=predict,
    inputs=gr.inputs.Textbox(lines=2, placeholder="Enter a movie review..."),
    outputs="text",
    title="IMDB Review Sentiment Analysis",
    description="A simple Gradio interface to predict sentiment of IMDB movie reviews using a RoBERTa model."
)

if __name__ == "__main__":
    iface.launch()