run_robotwin_train.sh

# export NCCL_SOCKET_IFNAME=bond0
export NCCL_IB_HCA=mlx5_2,mlx5_3
# export LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/stubs:$LIBRARY_PATH
# export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu/stubs:$LD_LIBRARY_PATH
# used for check save when communication
export NCCL_BLOCKING_WAIT=1
export NCCL_ASYNC_ERROR_HANDLING=1
export NCCL_TIMEOUT=1000  # timeout set to 1 hour (unit: seconds)
export WANDB_API_KEY='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'
###########################################################################################
# === Please modify the following paths according to your environment ===
Framework_name=RynnBrainOFT
freeze_module_list=''
base_vlm=playground/Pretrained_models/RynnBrain-CoP-8B
config_yaml=examples/Robotwin/train_files/starvla_cotrain_robotwin_abs.yaml
run_root_dir=./results/Checkpoints
data_mix=robotwin
run_id=0605_${data_mix}_RynnBrainOFT
# === End of environment variable configuration ===
###########################################################################################


# export WANDB_MODE=disabled

output_dir=${run_root_dir}/${run_id}
mkdir -p ${output_dir}
# mv this script to the output dir
cp $0 ${output_dir}/


accelerate launch \
  --config_file starVLA/config/deepseeds/deepspeed_zero2.yaml \
  --num_processes 8 \
  starVLA/training/train_starvla.py \
  --config_yaml ${config_yaml} \
  --framework.name ${Framework_name} \
  --framework.qwenvl.base_vlm ${base_vlm} \
  --datasets.vla_data.per_device_batch_size 4 \
  --datasets.vla_data.data_mix ${data_mix} \
  --trainer.freeze_modules ${freeze_module_list} \
  --trainer.max_train_steps 100000 \
  --trainer.save_interval 5000 \
  --trainer.logging_frequency 100 \
  --trainer.eval_interval 1000 \
  --run_root_dir ${run_root_dir} \
  --run_id ${run_id} \
  --wandb_project starVLA_Robotwin \
  --wandb_entity seramasumi-south-china-university-of-technology \
  # --is_debug True



##### Multi-Server Multi-GPU training script #####
  # accelerate launch \
  #   --config_file starVLA/config/deepseeds/deepspeed_zero2.yaml \
  #   --main_process_ip $MASTER_ADDR \
  #   --main_process_port $MASTER_PORT \
  #   --machine_rank $SLURM_PROCID \
  #   --num_machines $SLURM_NNODES \
  #   --num_processes=${TOTAL_GPUS} \
  #   starVLA/training/train_starvla.py \
  #   --config_yaml ${config_yaml} \
  #   --framework.name ${Framework_name} \
  #   --framework.qwenvl.base_vlm ${base_vlm} \
  #   --run_root_dir ${run_root_dir} \
  #   --run_id ${run_id} \
  #   --wandb_project your_project \
  #   --wandb_entity your_name
##### Multi-Server Multi-GPU training script #####