bgl3 · Sequence CNN

reference_bgl3_cnn_modulof3_20261104-161000

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	1.2863	0.6099	0.5795	0.3424	3467.0000	0.1566	4651.0000	1.0000	1.6952	2.9405	0.1000	0.6378	0.7800	1.0491	2.9405	0.0400	0.5033	0.7900	1.0117	2.9405	0.0600	0.5169	0.7879	1.0193	2.9405	0.0606	0.5191
validation	1.3131	0.1749	0.1855	0.0300	437.0000	0.0484	255.0000	0.7000	0.7015	2.3637	0.1000	0.0000	0.3840	0.4800	0.0343	2.3637	0.0800	0.0000	0.1805	0.4500	0.0105	2.3637	0.1400	0.0000	0.1887	0.6875	0.6978	2.3637	0.0625	0.0000	0.4026
test	1.3509	0.1442	0.1438	0.0203	437.0000	0.0185	532.0000	0.5000	0.0090	1.3581	0.0000	0.0000	0.1370	0.4800	-0.2086	1.5945	0.0000	0.0000	0.0834	0.4700	-0.1057	2.7610	0.0900	0.0000	0.1421	0.5000	-0.0651	1.3581	0.0000	0.0000	0.1186

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_modulof3
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: modulo
  split_fold: 3
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