bgl3 · Sequence CNN

reference_bgl3_cnn_contiguousf2_20261104-163909

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	1.8281	0.4130	0.4041	0.0207	4995.0000	0.0608	4830.0000	0.6000	0.3932	1.3581	0.2215	0.7000	0.6613	3.1160	0.0200	0.3452	0.6700	0.5162	3.1160	0.0200	0.3145	0.6471	0.4924	3.1160	0.0221	0.3138
validation	1.9425	-0.0354	-0.0357	-0.0029	1899.0000	0.0218	1938.0000	0.6000	0.2923	1.3581	0.1810	0.4000	-0.1877	2.0841	0.0000	0.0345	0.3700	-0.3555	2.0841	0.0200	-0.0281	0.4082	-0.1150	2.0841	0.0000	0.0607
test	2.1538	0.0660	0.0610	-0.0088	1465.0000	0.0378	952.0000	0.8000	0.7350	2.3637	0.3789	0.5400	0.0502	2.5419	0.0800	0.1902	0.5100	-0.0387	2.5419	0.0700	0.1703	0.5116	0.0071	2.5419	0.0930	0.1727

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_contiguousf2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: contiguous
  split_fold: 2
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