ds_zero2_config.json

{
  "train_batch_size": 16,
  "gradient_accumulation_steps": 1,
  "steps_per_print": 1,
  "optimizer": {
    "type": "Adam",
    "params": {
      "lr": 0.00015,
      "weight_decay": 1e-2
    }
  },
  "zero_optimization": {
    "stage": 2
  },
  "zero_allow_untested_optimizer": true,
  "gradient_clipping": 1.0,
  "fp16": {
    "enabled": true,
    "loss_scale": 0,
    "loss_scale_window": 1000,
    "hysteresis": 2,
    "min_loss_scale": 1
  },
  "activation_checkpointing": {
    "partition_activations": true,
    "contiguous_memory_optimization": false
  },
  "wall_clock_breakdown": false
}