ube4b · Sequence CNN

reference_ube4b_cnn_plm_randomf0_20261104-174824

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	2.0084	0.6888	0.6540	0.4736	12442.0000	0.4684	3.0000	1.0000	6.2463	9.1845	0.3000	0.8210	1.0000	5.4250	9.1845	0.3000	0.8541	1.0000	5.0145	9.1845	0.3800	0.8576	0.9949	3.9222	9.1845	0.5458	0.8821
validation	2.1590	0.6569	0.6181	0.4301	4135.0000	0.4414	319.0000	1.0000	5.1986	6.5758	0.1000	0.8176	1.0000	4.3376	6.5758	0.3600	0.8066	0.9900	4.1222	6.5758	0.4300	0.8403	0.9898	3.8307	6.5758	0.5330	0.8698
test	2.1569	0.6513	0.6162	0.4230	4174.0000	0.4162	142.0000	1.0000	4.9418	7.8259	0.1000	0.7521	1.0000	4.5483	8.0483	0.3600	0.8264	0.9900	4.1488	8.0483	0.4400	0.8419	0.9898	3.7989	8.6246	0.5685	0.8644

Loss per epoch

training validation

Config

name: reference_ube4b_cnn_plm_randomf0
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 0
model:
  name: cnn
  conv:
  - - 128
    - 3
  - - 128
    - 3
  - - 128
    - 3
  - - 128
    - 3
  - - 128
    - 3
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 128
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7