ube4b · Sequence CNN

reference_ube4b_cnn_plm_contiguousf2_20261104-184745

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	2.4137	0.4465	0.4456	0.1832	9957.0000	0.0801	14870.0000	1.0000	1.7537	3.2738	0.4026	0.9000	1.3286	3.8798	0.0000	0.4075	0.7700	0.8638	3.8798	0.0200	0.3629	0.5981	0.3937	5.5143	0.0771	0.3463
validation	3.5280	0.3061	0.3432	0.0601	323.0000	-0.0807	506.0000	0.4000	-0.1118	0.7848	0.2684	0.2800	-0.6021	1.4962	0.0000	0.2487	0.2900	-0.5507	3.8792	0.0500	0.2989	0.2800	-0.6442	0.7848	0.0000	0.2160
test	5.1683	0.3618	0.4100	0.1089	88.0000	0.0147	59.0000	0.6000	0.1788	1.6400	0.2817	0.3800	0.0422	4.5327	0.1600	0.3530	0.3200	-0.1843	5.7047	0.3600	0.4210	0.7500	-0.0658	0.1707	0.0000	0.2166

Loss per epoch

training validation

Config

name: reference_ube4b_cnn_plm_contiguousf2
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: contiguous
  split_fold: 2
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