pab1 · Fully connected

reference_pab1_fc_plm_modulof3_20261105-040121

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	precision_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	0.5529	0.9289	0.9209	0.8499	1256.0000	-0.0245	1475.0000	0.3000	0.0091	1.0896	0.1000	0.6643	0.3800	-0.0892	1.0896	0.0200	0.6951	0.3600	-0.0913	1.0896	0.0200	0.7355	0.3851	-0.0710	1.0896	0.0405	0.7656
validation	1.1350	0.8354	0.7489	0.6947	234.0000	0.0140	905.0000	0.3000	-0.2079	0.2554	0.0000	0.6445	0.3000	-0.2221	0.8367	0.0400	0.7011	0.3300	-0.1851	0.8367	0.1500	0.7552	0.2000	-0.2955	0.2554	0.0000	0.6192
test	1.2929	0.7983	0.7575	0.6301	223.0000	0.1341	509.0000	0.3000	-0.0924	0.6443	0.1000	0.6608	0.3600	-0.1779	0.7491	0.1600	0.7057	0.3400	-0.1800	0.7491	0.1900	0.7456	0.1875	-0.2661	0.6443	0.0625	0.6097

Loss per epoch

training validation

Config

name: reference_pab1_fc_plm_modulof3
task: dms
data:
  dataset: pab1
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: modulo
  split_fold: 3
model:
  name: fcn
  hidden:
  - 100
  - 100
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 128
  learning_rate: 0.001
  patience: 15
  min_delta: 1.0e-05
  seed: 7