dgx-spark-playbooks/nvidia/pytorch-fine-tune/assets/configs/config_fsdp_lora.yaml
2025-11-20 18:30:59 +00:00

29 lines
750 B
YAML

compute_environment: LOCAL_MACHINE
debug: false
distributed_type: FSDP
downcast_bf16: 'no'
enable_cpu_affinity: false
fsdp_config:
fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
fsdp_backward_prefetch: BACKWARD_PRE
fsdp_cpu_ram_efficient_loading: true
fsdp_forward_prefetch: false
fsdp_offload_params: false
fsdp_sharding_strategy: FULL_SHARD
fsdp_state_dict_type: SHARDED_STATE_DICT
fsdp_sync_module_states: true
fsdp_use_orig_params: true
machine_rank: 0
main_process_ip: < TODO: specify IP >
main_process_port: < TODO: specify port >
main_training_function: main
mixed_precision: 'bf16'
num_machines: 2
num_processes: 2
rdzv_backend: static
same_network: true
tpu_env: []
tpu_use_cluster: false
tpu_use_sudo: false
use_cpu: false