bgl3 · Graph CNN

reference_bgl3_gcn_mr5plm_modulof0_20261104-222312

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	0.7243	0.7541	0.7810	0.5625	109.0000	0.1025	606.0000	0.3000	-0.0751	0.2595	0.0000	0.6847	0.2800	-0.1272	0.3302	0.1600	0.7600	0.3100	-0.1226	0.7042	0.2900	0.8153	0.3448	-0.1034	0.3302	0.1034	0.7321
validation	0.7207	0.6022	0.6026	0.3615	47.0000	0.0032	60.0000	0.0000	-0.6269	-0.0020	0.0000	0.2322	0.1400	-0.4142	0.2216	0.1600	0.4942	0.1500	-0.3853	1.6908	0.2600	0.5763	0.0000	-0.2588	-0.0020	0.0000	0.4415
test	0.8254	0.5455	0.5961	0.2852	49.0000	-0.0284	184.0000	0.1000	-0.3012	0.3887	0.1000	0.5070	0.2000	-0.4534	0.4991	0.2000	0.4902	0.1900	-0.3795	0.4991	0.2900	0.6124	0.1429	-0.3505	0.3887	0.1429	0.4616

Loss per epoch

training validation

Config

name: reference_bgl3_gcn_mr5plm_modulof0
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  graph_threshold: 6
  split_scheme: modulo
  split_fold: 0
  min_reads: 5
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