defaults:
  learning_rate: 1e-5
  gradient_checkpointing: false
  gradient_accumulation_steps: 32
  per_device_train_batch_size: 2
  per_device_eval_batch_size: 2
  weight_decay: 0.00
  warmup_steps: 600
  eval_steps: 200
  save_steps: 500
  max_length: 512
  num_train_epochs: 3
  logging_steps: 10
  max_grad_norm: 2.0
  save_total_limit: 4
  eval_accumulation_steps:
  freeze_layer:
  datasets:
    - webgpt
  cache_dir: ~/.cache
  loss_fn: CrossEntropyLoss
  eval_size:
  log_dir: "base"

galactica-125:
  learning_rate: 5e-5
  model_name: facebook/galactica-125m
  weight_decay: 0.01
  warmup_steps: 600
  gradient_checkpointing: false
  gradient_accumulation_steps: 2
  per_device_train_batch_size: 4
  per_device_eval_batch_size: 4

debug:
  eval_steps: 20
  eval_size: 100