bgl3 · Sequence CNN

reference_bgl3_cnn_mr5_modulof2_20261104-222046

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	0.5737	0.8123	0.8109	0.6401	143.0000	0.1001	42.0000	0.7000	0.0108	0.4161	0.0000	0.6394	0.3400	-0.0376	1.6908	0.1800	0.7633	0.3100	-0.1300	1.6908	0.2200	0.7693	0.4074	-0.0016	0.9971	0.1481	0.7257
validation	0.8606	0.4067	0.3918	0.1418	44.0000	-0.2351	350.0000	0.4000	-0.4451	0.2595	0.1000	0.3431	0.1800	-0.4069	0.2595	0.2000	0.4764	0.1300	-0.4470	0.3192	0.2900	0.4837	0.2857	-0.6509	0.2595	0.0000	0.1810
test	1.0400	0.4028	0.4432	0.1148	23.0000	-0.2451	4.0000	0.2000	-0.7689	0.3940	0.1000	0.2453	0.1400	-0.4152	0.3940	0.3600	0.5955	0.0700	-0.5131	0.3940	0.3500	0.5515	0.1250	-0.8845	0.3940	0.1250	0.1575

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5_modulof2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index
  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