bgl3 · Graph CNN

reference_bgl3_gcn_mr5plm_randomf2_20261104-222611

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	0.5369	0.8490	0.8574	0.7117	124.0000	0.0549	4.0000	0.3000	0.0464	0.9971	0.2000	0.7327	0.2400	-0.0693	0.9971	0.1200	0.7935	0.2100	-0.1163	0.9971	0.2000	0.8238	0.2353	-0.0389	0.9971	0.1176	0.7777
validation	0.9800	0.6936	0.7202	0.4783	40.0000	-0.1253	200.0000	0.0000	-0.5790	-0.1056	0.0000	0.5454	0.0600	-0.3967	0.3365	0.1000	0.7160	0.1400	-0.3676	0.3365	0.3000	0.7755	0.0000	-0.5560	-0.1056	0.0000	0.5597
test	0.9351	0.7123	0.7368	0.5052	53.0000	-0.2683	25.0000	0.1000	-0.6850	0.0086	0.0000	0.4473	0.2000	-0.3205	1.6908	0.1800	0.7274	0.1300	-0.4620	1.6908	0.2600	0.6957	0.1818	-0.6204	0.0254	0.0000	0.4817

Loss per epoch

training validation

Config

name: reference_bgl3_gcn_mr5plm_randomf2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  graph_threshold: 6
  split_scheme: random
  split_fold: 2
  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