deepspeed_config.json

{
  "train_micro_batch_size_per_gpu": 32,
  "gradient_accumulation_steps": 4,
  "gradient_clipping": 1.0,
  "optimizer": {
    "type": "AdamW",
    "params": {
      "lr": 3e-4,
      "betas": [
        0.8,
        0.999
      ],
      "eps": 1e-8
    }
  },
  "scheduler": {
    "type": "WarmupLR",
    "params": {
      "warmup_min_lr": 0,
      "warmup_max_lr": 3e-4,
      "warmup_num_steps": 4000
    }
  },
  "zero_optimization": {
    "stage": 3,
    "overlap_comm": true,
    "allgather_partitions": true,
    "reduce_scatter": true,
    "allgather_bucket_size": 5e8,
    "reduce_bucket_size": 5e8,
    "contiguous_gradients": true,
    "offload_optimizer": {
      "device": "cpu",
      "pin_memory": true
    },
    "offload_param": {
      "device": "cpu",
      "pin_memory": true
    },
    "sub_group_size": 1e12
  },
  "activation_checkpointing": {
    "partition_activations": false,
    "contiguous_memory_optimization": false,
    "cpu_checkpointing": false,
    "number_checkpoints": null,
    "synchronize_checkpoint_boundary": false,
    "profile": false
  }
}