pab1 · Graph CNN

reference_pab1_gcn_plm_randomf2_20261104-183000

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	0.3520	0.9519	0.9422	0.9061	2664.0000	0.1298	7.0000	0.9000	0.4797	1.5204	0.1000	0.8000	0.8400	0.2790	1.5204	0.1400	0.7983	0.8800	0.2648	1.5204	0.1100	0.8253	0.8041	0.1926	1.5204	0.1551	0.8441
validation	0.4384	0.9391	0.9327	0.8818	917.0000	0.1060	95.0000	0.9000	0.3165	0.7658	0.0000	0.7828	0.7200	0.1812	0.7658	0.1400	0.8135	0.6900	0.1106	1.1615	0.1800	0.8234	0.6707	0.1346	0.7658	0.1707	0.8217
test	0.4431	0.9398	0.9324	0.8833	935.0000	0.0540	801.0000	0.6000	0.0931	0.8323	0.0000	0.6885	0.7000	0.1398	0.9945	0.1000	0.7799	0.6600	0.1089	0.9945	0.1000	0.8088	0.6707	0.1065	0.9945	0.0854	0.7967

Loss per epoch

training validation

Config

name: reference_pab1_gcn_plm_randomf2
task: dms
data:
  dataset: pab1
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  graph_threshold: 7
  split_scheme: random
  split_fold: 2
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