bgl3 · Sequence CNN

reference_bgl3_cnn_mr5plm_contiguousf2_20261104-222722

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	1.1209	0.6305	0.6636	0.3659	152.0000	-0.0489	962.0000	0.1000	-0.2255	0.1164	0.0000	0.6194	0.1400	-0.3912	0.1987	0.0200	0.6207	0.1800	-0.3805	0.2595	0.0900	0.6655	0.0625	-0.3770	0.1164	0.0000	0.6043
validation	1.3371	0.5279	0.5088	0.2600	38.0000	0.0124	227.0000	0.3000	-0.8535	0.3961	0.1000	0.3084	0.1200	-1.3526	0.3961	0.1200	0.0992	0.0900	-1.2322	0.7630	0.1500	0.1848	0.2727	-1.1318	0.3961	0.0909	0.1823
test	1.8104	0.3700	0.3150	0.0502	26.0000	-0.1262	681.0000	0.0000	-1.2895	-0.3805	0.0000	0.2237	0.0000	-1.9262	-0.1352	0.0000	-0.0955	0.0000	-1.7948	-0.1352	0.0400	-0.0248	0.0000	-1.3384	-0.3805	0.0000	0.2026

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_mr5plm_contiguousf2
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: contiguous
  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