bgl3 · Sequence CNN

reference_bgl3_cnn_mr5plm_randomf0_20261104-222215

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	0.5717	0.8343	0.8488	0.6956	134.0000	0.0003	122.0000	0.7000	0.0170	0.1135	0.0000	0.7248	0.3800	-0.0841	0.7042	0.1400	0.7963	0.3000	-0.1110	0.7042	0.2200	0.8332	0.3824	-0.0778	0.7042	0.1176	0.7705
validation	0.7710	0.7710	0.7940	0.5942	37.0000	0.1702	491.0000	0.2000	-0.0898	0.3347	0.2000	0.7574	0.1800	-0.2080	0.5578	0.2400	0.8130	0.1200	-0.2955	0.5578	0.3900	0.8173	0.1818	-0.0952	0.3347	0.1818	0.7588
test	0.8221	0.7420	0.7691	0.5472	46.0000	-0.0533	440.0000	0.2000	-0.6711	0.1579	0.0000	0.4355	0.1200	-0.3323	0.4991	0.1200	0.7139	0.1500	-0.3280	0.4991	0.3100	0.7688	0.1818	-0.6365	0.1579	0.0000	0.4574

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5plm_randomf0
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 0
  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