jd_train/ds_config_zero3.json

31 lines
719 B
JSON

{
"train_micro_batch_size_per_gpu": 1,
"gradient_accumulation_steps": 1,
"zero_optimization": {
"stage": 3,
"overlap_comm": true,
"contiguous_gradients": true,
"reduce_bucket_size": 80000000,
"stage3_prefetch_bucket_size": 40000000,
"stage3_param_persistence_threshold": 0,
"offload_optimizer": { "device": "none" },
"offload_param": { "device": "none" },
"stage3_gather_16bit_weights_on_model_save": false
},
"optimizer": {
"type": "AdamW",
"params": { "lr": 2e-5, "betas": [0.9, 0.999], "eps": 1e-8, "weight_decay": 0.1 }
},
"bf16": { "enabled": true },
"fp16": { "enabled": false },
"gradient_clipping": 1.0,
"wall_clock_breakdown": false
}