Protein Engineering

ube4b · Fully connected

reference_ube4b_fc_contiguousf0_20261104-145756

Metrics

Setmsepearsonrspearmanrr2above_wt_nabove_wt_spearmanrbudget_to_maxhit_rate_at_10mean_at_10max_at_10precision_at_10captured_at_10hit_rate_at_50mean_at_50max_at_50precision_at_50captured_at_50hit_rate_at_100mean_at_100max_at_100precision_at_100captured_at_100hit_rate_at_1pctmean_at_1pctmax_at_1pctprecision_at_1pctcaptured_at_1pct
train2.67180.62460.58660.36014241.00000.57181.00001.00005.92869.18450.30000.86590.98004.48789.18450.32000.79610.97003.95689.18450.34000.78220.94373.54479.18450.45070.7950
validation3.6169-0.1330-0.1233-0.0192458.0000-0.05571426.00000.0000-3.5648-1.71330.00000.0000-0.33960.0400-2.66522.87170.02000.0000-0.22410.0500-2.52882.87170.01000.0000-0.22040.0385-2.88552.87170.03850.0000-0.2460
test3.1417-0.0316-0.0167-0.38281218.0000-0.03622719.00000.2000-0.83431.29310.00000.0000-0.00350.1800-1.14941.56890.00000.0000-0.07960.1700-1.25951.56890.00000.0000-0.12210.1860-1.06391.56890.00000.0000-0.0576

Loss per epoch

training validation

Config

name: reference_ube4b_fc_contiguousf0
task: dms
data:
  dataset: ube4b
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: contiguous
  split_fold: 0
model:
  name: fcn
  hidden:
  - 100
  - 100
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7