description: angcb


# amlt cred storage set angcb4wjp --subscription e033d461-1923-44a7-872b-78f1d35a86dd --resource-group Shun  --allow-local-storage   False  

target:
  service: sing
  name: msrresrchlab
  workspace_name: epeastus

environment:
  registry: singularitybase.azurecr.io
  image: base/job/pytorch/acpt-2.2.1-py3.10-cuda12.1:20240312T225111416

storage:
  data:
    # storage_account_name: epeastus
    # container_name: cov19
    # mount_dir: /home/xiaofangui/Blob_EastUS
    storage_account_name: angcb4wjp
    container_name: angcb
    mount_dir: /data/angcb

code:
  # local directory of the code. this will be uploaded to the server.
  local_dir: $CONFIG_DIR/..


jobs:
  - name: ANGCB_land
    sku: 32G4-V100
    # process_count_per_node: 1
    # sla_tier: Standard
    sla_tier: premium
    priority: high
    identity: managed
    submit_args:
      env:
        SHARED_MEMORY_PERCENT: 1.0
        _AZUREML_SINGULARITY_JOB_UAI: "/subscriptions/e033d461-1923-44a7-872b-78f1d35a86dd/resourcegroups/Shun/providers/Microsoft.ManagedIdentity/userAssignedIdentities/epeastus_mi"
    command:
    - bash install_dep_pkgs.sh
    - pip list
    - free -h
    # - torchrun --nproc-per-node=4 multi_prediction_new.py --target ACCESS --exp_id 101 --model_type gobm

    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target CESM_ETHZ --exp_id 102 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target CNRM --exp_id 103 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target FESOM_REcoM --exp_id 104 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target IPSL --exp_id 105 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target MPIOM-HAMOCC --exp_id 106 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target MRI --exp_id 107 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target NEMO_PlankTOM --exp_id 108 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target NorESM --exp_id 109 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target Princeton --exp_id 110 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target CMEMS-LSCE-FFNN --exp_id 111 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target CSIR-ML6 --exp_id 112 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target JENA-MLS --exp_id 113 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target JMA-MLR --exp_id 114 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target LDEO-HPD --exp_id 115 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target NIES-ML3 --exp_id 116 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target OceanSODA-ETHZv2 --exp_id 117 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target UoEX-UEPFFNU --exp_id 118 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target VLIZ-SOMFFN --exp_id 119 --model_type dataproduct

    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target CNRM --exp_id 203 --model_type gobm
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target UoEX-UEPFFNU --exp_id 218 --model_type dataproduct
    # - torchrun --nproc-per-node=4 multi_prediction_new.py  --target average_gobm --exp_id 120 --model_type gobm
    # - CUDA_VISIBLE_DEVICES=4,5,6,7 torchrun --nproc-per-node=4 --master_port=12345 multi_prediction_new.py  --target average_datapro --exp_id 121 --model_type dataproduct