bgl3 · Sequence CNN

reference_bgl3_cnn_mr5_randomf1_20261104-221808

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	0.5021	0.8559	0.8595	0.7293	127.0000	0.0332	83.0000	0.2000	-0.1276	0.2388	0.0000	0.6616	0.1800	-0.1342	0.4506	0.0800	0.7697	0.2200	-0.1480	0.9971	0.1900	0.8122	0.2059	-0.1184	0.3302	0.0588	0.7497
validation	0.8698	0.7354	0.7593	0.5398	53.0000	-0.3277	172.0000	0.2000	-0.1882	0.1167	0.0000	0.6705	0.1600	-0.3164	0.3940	0.1600	0.7296	0.1200	-0.3404	0.3940	0.2800	0.7652	0.1818	-0.2351	0.1167	0.0000	0.6567
test	0.8452	0.7453	0.7681	0.5552	37.0000	0.0443	723.0000	0.1000	-0.4127	0.3251	0.1000	0.6048	0.1400	-0.3641	0.5578	0.1800	0.7276	0.1100	-0.3219	0.5578	0.2900	0.8018	0.1818	-0.3708	0.3251	0.0909	0.6279

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5_randomf1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: random
  split_fold: 1
  min_reads: 5
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