ube4b · Fully connected

reference_ube4b_fc_modulof1_20261104-153841

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	3.0898	0.4518	0.4674	0.0554	4756.0000	0.1557	4858.0000	0.8000	1.3867	4.1436	0.3796	0.6600	0.3924	4.1436	0.0200	0.2950	0.6000	0.2264	5.3569	0.0300	0.2935	0.5812	0.0764	5.3569	0.0641	0.3032
validation	4.8452	-0.1263	-0.1378	-0.0075	819.0000	-0.0989	219.0000	0.2000	-1.6825	0.7149	0.0000	-0.0771	0.2200	-1.5117	3.5441	0.0000	0.0000	-0.0621	0.1900	-1.5227	3.5441	0.0100	0.0000	-0.0708	0.1875	-1.4137	3.5441	0.0000	0.0000	-0.0439
test	3.1026	0.0555	0.0474	-0.0187	571.0000	0.0346	1592.0000	0.2000	-1.7399	0.1430	0.0000	-0.0495	0.1600	-1.3660	3.5623	0.0400	0.0000	0.0280	0.1800	-1.3785	3.5623	0.0400	0.0000	0.0286	0.2258	-1.3438	3.5623	0.0645	0.0000	0.0303

Loss per epoch

training validation

Config

name: reference_ube4b_fc_modulof1
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: modulo
  split_fold: 1
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