{
    "train_batch_size": 20,
    "gradient_accumulation_steps": 5,
    "bf16": {"enabled": true},
    "zero_optimization": {
        "stage": 2,
        "contiguous_gradients": true,
        "overlap_comm": true,
        "reduce_scatter": true,
        "reduce_bucket_size": 1e8,
        "allgather_bucket_size": 1e8,
    },
    "activation_checkpointing": {"partition_activations": true},
}