from transformers import TrainingArguments

training_args = TrainingArguments(
    output_dir="./results",
    num_train_epochs=3,
    per_device_train_batch_size=8,  # Optimized for T4 GPU
    per_device_eval_batch_size=8,
    warmup_steps=500,
    weight_decay=0.01,
    logging_dir='./logs',
    logging_steps=50,
    save_steps=500,
    eval_steps=500,
    evaluation_strategy="steps",
    save_strategy="steps",
    save_total_limit=3,
    load_best_model_at_end=True,
    learning_rate=5e-5,
    fp16=True,  # Using GPU features
    gradient_checkpointing=True
)