params.yaml

data:
  min_thermo_temp: 60.0
  min_temp_diff: 0.0
  max_meso_temp: 40.0
  min_align_cov: 0.95
  mmseq_params:
    coverage: 0.95
    min-seq-id: 0.5
    cluster-mode: 1
    similarity-type: 2
    sensitivity: 7
    max-seqs: 1000
    cluster-steps: 5
    cluster-reassign: 1
    e: 1e-3
  test_size: 0.1
  additional_filters:
    - 'abs({seq_len_diff})/{meso_seq_len} < 0.1'
  dev_sample_data: false
model:
  pretrained_model: "Rostlab/prot_t5_xl_uniref50"
  model_hyperparams:
    dropout_rate: 0.1
    relative_attention_max_distance: 250 
  task: 'translation' # one of 'reconstruction', 'translation'
  generation_max_length: 250
  generation_num_beams: 10
training:
  reweight: false # whether to weight the loss function by the number of sequences in each cluster
  freeze_early_layers: 0.2
  epochs:  0.5
  early_stopping: true
  early_stopping_patience: 4
  early_stopping_threshold: 0.01
  per_device_batch_size: 20
  auto_find_batch_size: false
  learning_rate: 1.e-4
  gradient_accumulation: 8
  gradient_checkpointing: true
  evals_per_save: 3 
  evals_per_epoch: 500 # null means save only at the end of training
  lr_scheduler_type: 'linear'
  warmup_ratio: 0.1
  label_smoothing_factor: 0.001
  optim: "adamw_hf"
  optim_args: null #"scale_parameter=False,relative_step=False"
  fp16: false
  bf16: true
  eval_single_example_per_cluster: true    # only during training
  dev_sample_data: false
testing:
  only_one_from_cluster: true
optimize:
  estimator: mAFminDGEstimator
  estimator_args: 
    af_params: ./.config/af_singularity_config.yaml
    use_relaxed: False
    num_replicates: 25
    fix_msas: true
    residue_length_norm: true
  n_trials: 100
  direction: minimize  # or 'maximize'
  sampler: NSGAIISampler
  cut_tails: null # number of gap spaces to keep on ends of the alignment
  gap_compressed_mutations: true # whether to consider a string of gaps a single mutation
  matrix: BLOSUM62
  match_score: 1
  mismatch_score: -1
  gapopen: -4
  gapextend: -1
  penalize_end_gaps: false
  sampler_args:
    population_size: 10
  optuna_overwrite: true