bgl3 · Sequence CNN

reference_bgl3_cnn_mr5_randomf2_20261104-222009

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	0.4707	0.8731	0.8738	0.7472	124.0000	0.1642	96.0000	0.6000	0.0213	0.3347	0.0000	0.7219	0.2400	-0.1213	0.5578	0.1600	0.7676	0.2200	-0.1183	0.9971	0.2100	0.8227	0.3235	-0.0762	0.5578	0.2059	0.7598
validation	0.9841	0.6930	0.7208	0.4762	40.0000	-0.1127	299.0000	0.0000	-0.4960	-0.1552	0.0000	0.5864	0.1200	-0.3733	0.3365	0.1200	0.7290	0.1200	-0.3793	0.3365	0.2700	0.7686	0.0000	-0.4927	-0.1552	0.0000	0.5912
test	0.8626	0.7413	0.7650	0.5436	53.0000	-0.1990	8.0000	0.3000	0.0585	1.6908	0.1000	0.7813	0.1800	-0.1695	1.6908	0.1400	0.8082	0.1700	-0.2101	1.6908	0.3400	0.8395	0.3636	0.0555	1.6908	0.0909	0.7887

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5_randomf2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: random
  split_fold: 2
  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