gb1 · Sequence CNN

reference_gb1_cnn_plm_contiguousf0_20261104-175456

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	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	0.1744	0.9856	0.9706	0.9702	32275.0000	0.9373	1669.0000	1.0000	1.6885	1.9362	0.9051	1.0000	1.7383	2.1952	0.1400	0.9519	1.0000	1.6928	2.1952	0.2100	0.9594	1.0000	1.4244	2.4539	0.8126	0.9923
validation	2.4855	-0.1172	-0.1184	-0.2317	5403.0000	0.0174	1168.0000	0.0000	-1.2601	-0.7863	-0.1044	0.1400	-1.2327	0.9431	0.0000	-0.1003	0.1500	-1.3948	0.9431	0.0000	-0.1815	0.2139	-1.2313	1.0188	0.0149	-0.1087
test	3.1827	0.3802	0.3908	-0.0368	1866.0000	0.1492	7217.0000	0.9000	0.3801	0.7502	0.7826	0.8400	0.2792	0.7876	0.1600	0.8129	0.6500	0.0321	0.9352	0.1700	0.7460	0.5622	-0.1231	0.9352	0.1493	0.7103

Loss per epoch

training validation

Config

name: reference_gb1_cnn_plm_contiguousf0
task: dms
data:
  dataset: gb1
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: contiguous
  split_fold: 0
model:
  name: cnn
  conv:
  - - 128
    - 17
  - - 128
    - 17
  - - 128
    - 17
  hidden:
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 32
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7