bgl3 · Sequence CNN

reference_bgl3_cnn_plm_modulof1_20261104-181129

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	1.5761	0.3650	0.3350	0.1172	3546.0000	-0.0219	2440.0000	0.5000	0.2870	1.3581	0.1904	0.6400	0.4667	2.2054	0.0000	0.2800	0.6200	0.5163	3.0445	0.0800	0.3171	0.6263	0.5229	3.0445	0.0808	0.3191
validation	1.3882	0.3178	0.3176	0.0994	432.0000	-0.0741	624.0000	0.5000	0.3883	1.4833	0.3235	0.3600	-0.0374	2.0174	0.0200	0.2022	0.3800	-0.0038	2.2054	0.0800	0.2374	0.4118	0.2052	1.5218	0.0000	0.2687
test	1.3542	0.3737	0.3141	0.1349	424.0000	-0.1124	475.0000	0.5000	0.1923	1.3581	0.2172	0.4800	0.1470	2.0841	0.0400	0.2366	0.3600	-0.0205	2.0841	0.1000	0.1899	0.4667	0.2063	1.4833	0.0000	0.2291

Loss per epoch

training validation

Config

name: reference_bgl3_cnn_plm_modulof1
task: dms
data:
  dataset: bgl3
  encoding: one_hot,aa_index,plm,plm_sum
  split_dir: null
  split_scheme: modulo
  split_fold: 1
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