defaults:
  - base_pytorch
  - _self_
  
tasks: [training]

training:
  lr: 1e-5
  precision: bf16-mixed
  batch_size: 1
  max_epochs: -1
  max_steps: 10000000
  checkpointing:
    every_n_train_steps: 2000
    every_n_epochs: null
    save_weights_only: true
    filename: "latest"
  optim:
    accumulate_grad_batches: 4
    gradient_clip_val: null
  data:
      num_workers: 5 # number of CPU threads for data preprocessing.

validation:
  precision: bf16-mixed
  val_every_n_step: 1000
  val_every_n_epoch: null
  batch_size: 1
  limit_batch: 1
  data:
    num_workers: 1 # number of CPU threads for data preprocessing, for validation.

test:
  precision: bf16-mixed
  limit_batch: null
  batch_size: 1
  data:
    num_workers: 1 # number of CPU threads for data preprocessing, for test.

find_unused_parameters: False