gb1 · Fully connected

reference_gb1_fc_contiguousf3_20261104-161836

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	0.5752	0.9470	0.9617	0.8820	30012.0000	0.7399	87.0000	1.0000	1.6952	1.8527	0.8933	1.0000	1.6132	1.8991	0.1600	0.9246	1.0000	1.6493	2.5295	0.3100	0.9545	0.9995	1.3152	2.5295	0.7295	0.9785
validation	4.4096	0.1463	0.1364	0.0057	5856.0000	0.1450	5253.0000	0.3000	-0.4288	0.7253	0.0000	0.3902	0.4800	-0.7015	0.7253	0.0000	0.0000	0.3093	0.4900	-0.8268	0.8949	0.0000	0.0000	0.2707	0.4925	-0.6975	0.8949	0.0000	0.0000	0.3288
test	10.4658	-0.1991	-0.2208	-0.9082	2473.0000	0.0494	9589.0000	0.0000	-5.7722	-5.0631	0.0000	-0.4710	0.0000	-5.5067	-3.5964	0.0000	0.0000	-0.4251	0.0000	-5.4892	-3.5283	0.0000	0.0000	-0.4273	0.0050	-5.4607	0.7301	0.0000	0.0000	-0.4299

Loss per epoch

training validation

Config

name: reference_gb1_fc_contiguousf3
task: dms
data:
  dataset: gb1
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: contiguous
  split_fold: 3
model:
  name: fcn
  hidden:
  - 1000
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7