bgl3 · Sequence CNN

reference_bgl3_cnn_mr5plm_modulof2_20261104-222640

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	0.5133	0.8372	0.8456	0.6779	143.0000	0.1133	91.0000	0.5000	-0.0563	0.3347	0.0000	0.6079	0.3600	-0.0361	0.7042	0.1800	0.7641	0.2800	-0.0938	1.6908	0.2100	0.7922	0.4815	0.0025	0.7042	0.1852	0.7279
validation	0.6646	0.5875	0.5748	0.3373	44.0000	-0.2016	300.0000	0.3000	-0.3879	0.1195	0.0000	0.3857	0.2000	-0.3439	0.1254	0.2000	0.5366	0.1800	-0.3656	0.3961	0.4000	0.5695	0.4286	-0.1484	0.1195	0.0000	0.5380
test	0.8537	0.5322	0.6021	0.2734	23.0000	0.0766	8.0000	0.1000	-0.9286	0.3940	0.1000	0.1288	0.1200	-0.4111	0.3940	0.2600	0.5991	0.1300	-0.4603	0.3940	0.3700	0.6006	0.1250	-0.8072	0.3940	0.1250	0.2127

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5plm_modulof2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: modulo
  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