bgl3 · Fully connected

reference_bgl3_fc_contiguousf1_20261104-154008

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	1.8929	0.5144	0.4942	0.0324	5657.0000	0.1074	8087.0000	1.0000	1.4189	3.1160	0.1000	0.5080	0.8000	0.6979	3.1160	0.0200	0.3412	0.7900	0.6746	3.1160	0.0100	0.3553	0.7267	0.5943	3.1160	0.0333	0.3400
validation	2.1365	-0.0841	-0.0819	-0.0007	1465.0000	-0.0581	429.0000	0.2000	-1.5797	1.2150	0.0000	0.0000	-0.3345	0.2000	-1.5204	1.5945	0.0000	0.0000	-0.3583	0.2600	-1.2429	2.2054	0.0100	0.0000	-0.2798	0.2326	-1.4271	1.5945	0.0000	0.0000	-0.3211
test	1.6792	-0.0346	-0.0521	-0.0137	1203.0000	-0.0131	3014.0000	0.4000	0.0307	1.6946	0.0000	0.0000	0.1278	0.4400	-0.0410	2.1466	0.0200	0.0000	0.1112	0.3800	-0.2691	2.1466	0.0100	0.0000	0.0222	0.4412	-0.1012	2.1466	0.0294	0.0000	0.0847

Loss per epoch

training validation

Config

name: reference_bgl3_fc_contiguousf1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: contiguous
  split_fold: 1
model:
  name: fcn
  hidden:
  - 100
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 32
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7