pab1 · Sequence CNN

reference_pab1_cnn_plm_contiguousf4_20261105-043736

Metrics

Set	mse	pearsonr	spearmanr	r2	above_wt_n	above_wt_spearmanr	budget_to_max	hit_rate_at_10	mean_at_10	max_at_10	captured_at_10	hit_rate_at_50	mean_at_50	max_at_50	precision_at_50	captured_at_50	hit_rate_at_100	mean_at_100	max_at_100	precision_at_100	captured_at_100	hit_rate_at_1pct	mean_at_1pct	max_at_1pct	precision_at_1pct	captured_at_1pct
train	0.5147	0.9407	0.9271	0.8761	1652.0000	0.0790	802.0000	0.7000	0.0920	0.4835	0.7406	0.7200	0.1044	0.8395	0.0200	0.8017	0.6500	0.0551	0.8395	0.0600	0.8125	0.5778	0.0021	0.8395	0.1111	0.8177
validation	1.7741	0.7299	0.7300	0.5214	222.0000	-0.0232	2117.0000	0.2000	-0.2863	0.0634	0.7514	0.1800	-0.3092	0.3723	0.0400	0.7803	0.1800	-0.3921	0.3723	0.0600	0.7764	0.1633	-0.3231	0.2373	0.0204	0.7748
test	1.4206	0.7506	0.6155	0.2384	931.0000	0.0352	1380.0000	0.2000	-0.2098	0.2331	0.3911	0.2000	-0.3365	0.8601	0.0400	0.3794	0.2400	-0.3259	0.8601	0.0300	0.4162	0.2222	-0.3353	0.8601	0.0444	0.3760

Loss per epoch

training validation

Config

name: reference_pab1_cnn_plm_contiguousf4
task: dms
data:
  dataset: pab1
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: contiguous
  split_fold: 4
model:
  name: cnn
  conv:
  - - 128
    - 17
  - - 128
    - 17
  - - 128
    - 17
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 128
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7