bgl3 · Graph CNN

reference_bgl3_gcn_contiguousf3_20261104-161756

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	1.9241	0.4936	0.4878	0.0243	4815.0000	0.1127	1560.0000	0.9000	1.1334	1.6946	0.4566	0.7000	0.7484	2.0841	0.0000	0.3918	0.7000	0.6766	2.2610	0.0100	0.3897	0.6957	0.6413	2.2610	0.0145	0.3883
validation	1.7777	-0.0720	-0.0512	-0.0010	1997.0000	0.0158	4690.0000	0.4000	-0.5812	1.2150	0.0000	-0.1323	0.2800	-0.5988	2.0174	0.0000	0.0000	-0.1590	0.3500	-0.3810	2.3637	0.0500	0.0000	-0.0796	0.2766	-0.6430	1.9459	0.0000	0.0000	-0.1752
test	1.9430	-0.0860	-0.0432	-0.0032	1899.0000	-0.0292	2795.0000	0.0000	-2.2628	-0.8703	0.0000	-0.6231	0.0200	-2.2532	0.0852	0.0000	0.0000	-0.7136	0.1600	-1.6487	1.7856	0.0000	0.0000	-0.5286	0.0204	-2.2492	0.0852	0.0000	0.0000	-0.7110

Loss per epoch

training validation

Config

name: reference_bgl3_gcn_contiguousf3
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  split_dir: null
  graph_threshold: 6
  split_scheme: contiguous
  split_fold: 3
model:
  name: gcn
  filters:
  - 32
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 128
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7