bgl3 · Sequence CNN

reference_bgl3_cnn_mr5plm_randomf1_20261104-222404

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	0.5046	0.8583	0.8651	0.7279	127.0000	0.0085	202.0000	0.4000	-0.0739	0.0463	0.0000	0.6848	0.2800	-0.1237	0.3887	0.0600	0.7750	0.2200	-0.1397	0.5351	0.1600	0.8166	0.3529	-0.0736	0.3887	0.0294	0.7715
validation	0.8716	0.7355	0.7653	0.5388	53.0000	-0.1303	79.0000	0.2000	-0.1724	0.1090	0.0000	0.6776	0.2400	-0.1671	0.7042	0.2200	0.8095	0.2500	-0.1973	1.6908	0.3900	0.8468	0.2727	-0.1386	0.1987	0.0000	0.7005
test	0.7965	0.7651	0.7877	0.5808	37.0000	0.1387	592.0000	0.3000	-0.0833	0.3347	0.2000	0.7605	0.1200	-0.2185	0.4161	0.1600	0.8072	0.1200	-0.2569	0.5578	0.3700	0.8399	0.2727	-0.0806	0.3347	0.1818	0.7658

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5plm_randomf1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 1
  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