bgl3 · Sequence CNN

reference_bgl3_cnn_contiguousf3_20261104-161751

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	1.6223	0.4988	0.4778	0.1773	4815.0000	0.1116	4351.0000	0.6000	0.2312	1.4833	0.1952	0.5400	0.3960	2.0841	0.0000	0.2759	0.6000	0.5265	2.6204	0.0400	0.3374	0.6449	0.5877	2.6204	0.0435	0.3691
validation	1.7835	-0.0947	-0.0722	-0.0042	1997.0000	0.0238	531.0000	0.3000	-0.8147	1.6946	0.0000	-0.2111	0.3000	-1.1140	1.6946	0.0000	0.0000	-0.3588	0.3000	-0.9528	2.8548	0.0100	0.0000	-0.3166	0.2766	-1.2372	1.6946	0.0000	0.0000	-0.4044
test	1.9438	-0.0734	-0.0634	-0.0036	1899.0000	0.0221	509.0000	0.4000	-1.1966	1.4833	0.0000	-0.2876	0.4200	-0.7506	1.8419	0.0000	0.0000	-0.1694	0.3600	-0.9012	2.0841	0.0100	0.0000	-0.2393	0.4082	-0.7873	1.8419	0.0000	0.0000	-0.1824

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_contiguousf3
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: contiguous
  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