You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

config3.yaml 1.1KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253
  1. default: &default
  2. use_tqdm: true
  3. random_seed: 42
  4. base_save_path: /home/msadraei/trained_final
  5. model_name: t5-small
  6. project_name_prefix: iclr_orig_t5
  7. experiment_name_suffix: null
  8. train_batch_size: 24
  9. valid_batch_size: 24
  10. remove_dropout: true
  11. learning_rate: 0.01
  12. weight_decay: 0.01
  13. num_epochs: 80
  14. peft_params: null # no mutation
  15. hot_modules:
  16. - sadcl
  17. best_finder:
  18. save: True
  19. metric: valid_mean
  20. higher_better: true
  21. tasks:
  22. - superglue:rte
  23. - superglue:cb
  24. - superglue:wic
  25. - superglue:copa
  26. - glue:cola
  27. - glue:mrpc
  28. - superglue:boolq
  29. - glue:qqp
  30. - glue:qnli
  31. - glue:mnli
  32. - glue:sst2
  33. - glue:stsb
  34. pp: &pp
  35. - /home/msadraei/trained_final/hzi_cluster_t5_small_glue-mnli/10_combine_128
  36. - /home/msadraei/trained_final/hzi_cluster_t5_small_glue-sst2/10_combine_128
  37. - /home/msadraei/trained_final/hzi_cluster_t5_small_glue-qqp/10_combine_128
  38. run_configs:
  39. - <<: *default
  40. peft_params:
  41. kind: combine
  42. n_tokens: 10
  43. n_comb_tokens: 128
  44. # - <<: *default
  45. # learning_rate: 0.3
  46. # peft_params:
  47. # kind: residual
  48. # n_tokens: 10
  49. # mlp_size: 128