{ "train_batch_size": "auto", "train_micro_batch_size_per_gpu": "auto", "gradient_accumulation_steps": "auto", "gradient_clipping": 1.0, "zero_optimization": { "stage": 2, "allgather_partitions": true, "allgather_bucket_size": 100000000, "reduce_scatter": true, "reduce_bucket_size": 100000000, "overlap_comm": true, "contiguous_gradients": true }, "fp16": { "enabled": true, "loss_scale": 0, "loss_scale_window": 1000, "initial_scale_power": 16, "hysteresis": 2, "min_loss_scale": 1 }, "bf16": { "enabled": false }, "optimizer": { "type": "AdamW", "params": { "lr": "auto", "betas": [0.9, 0.95], "eps": 1e-8, "weight_decay": 0.1 } }, "scheduler": { "type": "WarmupDecayLR", "params": { "warmup_min_lr": 0, "warmup_max_lr": "auto", "warmup_num_steps": "auto", "total_num_steps": "auto" } }, "steps_per_print": 1000, "wall_clock_breakdown": false }