bgl3 · Sequence CNN

reference_bgl3_cnn_mr5plm_randomf2_20261104-222554

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	0.4927	0.8652	0.8726	0.7354	124.0000	0.0795	80.0000	0.4000	-0.0210	0.3887	0.7038	0.2400	-0.1096	0.5578	0.1600	0.7735	0.2500	-0.1070	0.9971	0.2200	0.8288	0.3235	-0.1010	0.5578	0.1765	0.7479
validation	0.9085	0.7193	0.7449	0.5164	40.0000	-0.1217	198.0000	0.3000	-0.5812	0.0874	0.5443	0.2200	-0.3410	0.3282	0.2800	0.7469	0.1700	-0.3327	0.3517	0.4100	0.7961	0.2727	-0.5508	0.0874	0.0000	0.5623
test	0.8544	0.7410	0.7694	0.5480	53.0000	-0.2174	72.0000	0.3000	-0.2229	0.2595	0.6549	0.2800	-0.2444	0.2595	0.2400	0.7681	0.2300	-0.2595	1.6908	0.4000	0.8113	0.2727	-0.2444	0.2595	0.0000	0.6524

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5plm_randomf2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 2
  min_reads: 5
model:
  name: cnn
  conv:
  - - 32
    - 17
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7