bgl3 · Sequence CNN

reference_bgl3_cnn_plm_contiguousf1_20261104-181600

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	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	1.7050	0.3718	0.3468	0.1285	5657.0000	-0.0146	4053.0000	0.7000	0.6618	2.2054	0.2906	0.5800	0.4103	2.3637	0.2476	0.6300	0.5053	2.3637	0.0100	0.2967	0.6200	0.4443	2.6204	0.0333	0.2861
validation	2.0567	0.2120	0.0993	0.0367	1465.0000	-0.0366	433.0000	0.6000	0.6407	1.6946	0.3498	0.2400	-0.8009	1.6946	-0.1070	0.3300	-0.5731	1.9459	0.0200	-0.0294	0.2791	-0.6984	1.6946	0.0000	-0.0702
test	1.5378	0.3312	0.2849	0.0716	1203.0000	-0.0012	1862.0000	0.5000	0.0855	1.5945	0.1477	0.5600	0.2724	1.6946	0.2356	0.4500	0.1762	2.2054	0.0200	0.2120	0.5588	0.1864	1.5945	0.0000	0.1954

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_plm_contiguousf1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: contiguous
  split_fold: 1
model:
  name: cnn
  conv:
  - - 32
    - 17
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7