{
    "train_micro_batch_size_per_gpu": 16,
    "gradient_accumulation_steps": 2,
    "steps_per_print": 100,
    "optimizer": {
        "type": "Adam",
        "params": {
            "lr": 1e-4
        }
    },
    "scheduler": {
        "type": "WarmupDecayLR",
        "params": {
            "total_num_steps": 10000000,
            "warmup_max_lr": 0.0001,
            "warmup_num_steps": 1000
        }
    },
    "fp16": {
        "enabled": false,
        "loss_scale": 0,
        "loss_scale_window": 1000,
        "initial_scale_power": 16,
        "hysteresis": 2,
        "min_loss_scale": 1
    },
    "bf16": {
        "enabled": true
    },
    "zero_optimization": {
        "stage": 3,
        "offload_optimizer": {
            "device": "cpu",
            "pin_memory": true
        },
        "offload_param": {
            "device": "cpu",
            "pin_memory": true
        },
        "overlap_comm": true,
        "contiguous_gradients": true,
        "sub_group_size": 1e9,
        "reduce_bucket_size": "auto",
        "stage3_prefetch_bucket_size": "auto",
        "stage3_param_persistence_threshold": "auto",
        "stage3_max_live_parameters": 1e9,
        "stage3_max_reuse_distance": 1e9,
        "stage3_gather_16bit_weights_on_model_save": true
    },
    "activation_checkpointing": {
        "partition_activations": true,
        "cpu_checkpointing": true,
        "contiguous_memory_optimization": true,
        "synchronize_checkpoint_boundary": false,
        "profile": false
    }
}