pab1 · Graph CNN

reference_pab1_gcn_contiguousf3_20261104-162323

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	0.8178	0.8977	0.9108	0.7939	1489.0000	0.1105	394.0000	0.5000	-0.0994	0.6496	0.1000	0.7286	0.4600	-0.0895	0.6496	0.0200	0.7753	0.4600	-0.0587	0.6496	0.0500	0.8077	0.4579	-0.0536	0.6496	0.0794	0.8348
validation	2.1863	-0.1789	-0.1474	-0.1721	931.0000	-0.0017	3488.0000	0.0000	-4.1027	-2.4547	0.0000	0.0000	-1.5282	0.0200	-2.9147	0.4984	0.0000	0.0000	-1.0881	0.0300	-2.7386	0.4984	0.0100	0.0000	-1.0668	0.0000	-3.0072	-0.1221	0.0000	0.0000	-1.1286
test	2.7693	-0.0639	-0.1129	-0.0913	383.0000	-0.0411	2027.0000	0.1000	-2.6861	0.1586	0.0000	0.0000	-0.7460	0.0600	-2.8591	0.1586	0.0000	0.0000	-0.9567	0.0500	-2.8424	0.1586	0.0000	0.0000	-1.0257	0.0312	-2.5789	0.1586	0.0000	0.0000	-0.7628

Loss per epoch

training validation

Config

name: reference_pab1_gcn_contiguousf3
task: dms
data:
  dataset: pab1
  encoding: one_hot,aa_index
  split_dir: null
  graph_threshold: 7
  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