| { | |
| "bf16": { | |
| "enabled": true | |
| }, | |
| "train_micro_batch_size_per_gpu": "auto", | |
| "train_batch_size": "auto", | |
| "gradient_accumulation_steps": "auto", | |
| "zero_optimization": { | |
| "stage": 3, | |
| "overlap_comm": true, | |
| "contiguous_gradients": true, | |
| "reduce_scatter": true, | |
| "stage3_max_live_parameters": 3000000000.0, | |
| "stage3_max_reuse_distance": 3000000000.0, | |
| "stage3_prefetch_bucket_size": 500000000.0, | |
| "stage3_param_persistence_threshold": 1000000.0, | |
| "sub_group_size": 1000000000.0, | |
| "offload_optimizer": { | |
| "device": "cpu", | |
| "pin_memory": true | |
| }, | |
| "offload_param": { | |
| "device": "cpu", | |
| "pin_memory": true | |
| } | |
| } | |
| } |