bgl3 · Fully connected

reference_bgl3_fc_mr5plm_randomf1_20261104-222406

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	0.5123	0.8534	0.8601	0.7238	127.0000	0.0530	18.0000	0.4000	-0.0081	0.3302	0.0000	0.7133	0.3000	-0.0728	0.9971	0.1200	0.8006	0.3200	-0.0911	0.9971	0.2400	0.8428	0.3235	-0.0287	0.9971	0.1176	0.7933
validation	0.9067	0.7224	0.7425	0.5203	53.0000	0.0485	108.0000	0.1000	-0.3833	0.1135	0.0000	0.5828	0.1800	-0.4095	0.2595	0.1800	0.6798	0.2100	-0.3450	0.7042	0.3600	0.7625	0.0909	-0.6892	0.1135	0.0000	0.4504
test	0.8432	0.7489	0.7609	0.5562	37.0000	0.0301	524.0000	0.1000	-0.3806	0.3347	0.1000	0.6200	0.1400	-0.3558	0.3347	0.1400	0.7321	0.1300	-0.3273	0.4161	0.3100	0.7987	0.0909	-0.3738	0.3347	0.0909	0.6265

Loss per epoch

training validation

Config

name: reference_bgl3_fc_mr5plm_randomf1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: random
  split_fold: 1
  min_reads: 5
model:
  name: fcn
  hidden:
  - 100
  - 100
  dropout: 0.2
train:
  epochs: 300
  batch_size: 32
  learning_rate: 0.0001
  patience: 15
  min_delta: 1.0e-05
  seed: 7