Protein Engineering

gb1 · Fully connected

reference_gb1_fc_modulof1_20261104-153430

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
train0.58890.96240.96670.885135035.00000.7807152.00001.00001.58171.74340.00000.85321.00001.62572.09420.08000.91311.00001.60212.09420.17000.93090.99951.36232.52950.69390.9772
validation5.4354-0.2709-0.2461-0.22411545.00000.07957027.00000.0000-5.4091-4.33480.00000.0000-0.70680.0000-5.5491-4.00260.00000.0000-0.77630.0000-5.4865-1.73110.00000.0000-0.77370.0000-5.1443-0.27170.00000.0000-0.6916
test6.3363-0.2863-0.2519-0.09274408.00000.070216266.00000.0000-5.9766-5.28590.00000.0000-0.92840.0000-5.8588-4.67060.00000.0000-0.94690.0000-5.6778-3.96630.00000.0000-0.92030.0000-5.3151-1.17440.00000.0000-0.8414

Loss per epoch

training validation

Config

name: reference_gb1_fc_modulof1
task: dms
data:
  dataset: gb1
  encoding: one_hot,aa_index
  split_dir: null
  split_scheme: modulo
  split_fold: 1
model:
  name: fcn
  hidden:
  - 1000
  dropout: 0.2
train:
  epochs: 300
  batch_size: 64
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7