ube4b · Fully connected

reference_ube4b_fc_plm_randomf3_20261105-035641

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	2.0737	0.6727	0.6336	0.4518	12385.0000	0.4408	3.0000	1.0000	5.7413	9.1845	0.1000	0.7566	1.0000	5.3908	9.1845	0.2600	0.8426	1.0000	5.0605	9.1845	0.4000	0.8595	0.9898	3.9599	9.1845	0.5729	0.8869
validation	2.1604	0.6621	0.6247	0.4381	4163.0000	0.4521	59.0000	1.0000	4.7082	7.7437	0.2000	0.7348	1.0000	4.6755	7.7437	0.3600	0.8482	1.0000	4.3649	9.1537	0.4500	0.8712	0.9848	3.8512	9.1537	0.5330	0.8681
test	2.1130	0.6643	0.6277	0.4411	4203.0000	0.4333	31.0000	1.0000	5.2975	6.7153	0.3000	0.8479	1.0000	4.6626	6.9328	0.3800	0.8632	1.0000	4.4256	6.9328	0.5400	0.8979	0.9949	3.8139	6.9328	0.5330	0.8759

Loss per epoch

training validation

Config

name: reference_ube4b_fc_plm_randomf3
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 3
model:
  name: fcn
  hidden:
  - 100
  - 100
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7