ube4b · Graph CNN

reference_ube4b_gcn_plm_randomf1_20261104-180845

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	1.9482	0.6982	0.6652	0.4857	12540.0000	0.4678	55.0000	1.0000	5.8116	7.7437	0.2000	0.7984	1.0000	5.2362	7.8259	0.2400	0.8419	1.0000	5.0562	9.1537	0.3300	0.8675	1.0000	3.8937	9.1537	0.5424	0.8813
validation	2.1748	0.6563	0.6178	0.4307	4076.0000	0.4379	3.0000	1.0000	5.6328	9.1845	0.1000	0.7955	1.0000	4.7058	9.1845	0.4000	0.8395	1.0000	4.4429	9.1845	0.4100	0.8734	0.9949	3.9024	9.1845	0.5330	0.8701
test	2.1498	0.6578	0.6188	0.4325	4135.0000	0.4478	134.0000	1.0000	5.3931	6.5758	0.1000	0.8416	1.0000	4.2944	6.5758	0.3600	0.8006	1.0000	4.0704	6.5758	0.3900	0.8324	0.9797	3.7450	7.6288	0.4924	0.8556

Loss per epoch

training validation

Config

name: reference_ube4b_gcn_plm_randomf1
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  graph_threshold: 7
  split_scheme: random
  split_fold: 1
model:
  name: gcn
  filters:
  - 128
  - 128
  - 128
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 128
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7