pab1 · Graph CNN

reference_pab1_gcn_plm_contiguousf2_20261104-184733

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	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	0.5795	0.9220	0.9209	0.8414	1747.0000	0.0558	1002.0000	0.6000	-0.0424	0.5640	0.6507	0.5200	0.0312	0.5640	0.0000	0.7367	0.5400	0.0208	0.8323	0.0600	0.7647	0.5722	0.0359	0.8601	0.0722	0.7995
validation	0.8522	0.8190	0.7052	0.6642	383.0000	-0.0445	1704.0000	0.0000	-0.5116	-0.2740	0.3409	0.1000	-0.3435	0.4974	0.0400	0.4884	0.1900	-0.2958	0.4974	0.0700	0.5586	0.0312	-0.3623	0.0138	0.0000	0.4579
test	1.7416	0.7844	0.8019	0.5900	475.0000	-0.0464	1065.0000	0.4000	-0.0364	0.5226	0.7609	0.2200	-0.1149	0.5226	0.0600	0.7890	0.2300	-0.1328	0.5226	0.0400	0.8103	0.2083	-0.1191	0.5226	0.0625	0.7859

Loss per epoch

training validation

Config

name: reference_pab1_gcn_plm_contiguousf2
task: dms
data:
  dataset: pab1
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  graph_threshold: 7
  split_scheme: contiguous
  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