SuperposPrompt_Thesis/09_Cluster/run_hyperparam_effect/config1.yaml

default: &default
  use_tqdm: true 
  random_seed: 42
  base_save_path: /home/msadraei/trained_final
  model_name: google/t5-base-lm-adapt
  project_name_prefix: hzi_cluster_comp_run
  experiment_name_suffix: null
  train_batch_size: 32
  valid_batch_size: 32
  remove_dropout: true
  learning_rate: 0.01
  weight_decay: 0.01
  num_epochs: 80
  peft_params: null  # no mutation
  hot_modules:
  - sadcl
  - classifier
  best_finder:
    save: True
    metric: valid_mean
    higher_better: true
  tasks:
  - glue:mrpc
  - glue:cola


run_configs:
- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 4

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 8

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 16

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 32

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 64

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 128

- <<: *default
  peft_params:
    kind: combine
    n_tokens: 10
    n_comb_tokens: 256