From: Prediction of chemical reaction yields with large-scale multi-view pre-training
Dataset | Split type | # Training | # Test | Out-of-sample type |
---|---|---|---|---|
Buchwald-Hartwig | Test 1 | 3057 | 898 | Ligand-based |
(3955 reactions) | Test 2 | 3055 | 900 | Ligand-based |
 | Test 3 | 3058 | 897 | Ligand-based |
 | Test 4 | 3055 | 900 | Ligand-based |
 | Plate 1 | 2880 | 1075 | Ligand-based |
 | Plate 2 | 2515 | 1440 | Lligand-based |
 | Plate 3 | 2515 | 1440 | Ligand-based |
 | Plate 2 new | 2515 | 1440 | Ligand-based |
 | Halide Br | 2636 | 1319 | Reactant-based |
 | Halide Cl | 2637 | 1318 | Reactant-based |
 | Halide I | 2637 | 1318 | reactant-based |
 | Pyridyl | 2372 | 1583 | Reactant-based |
 | Nonpyridyl | 1583 | 2372 | Reactant-based |
 | random | 2768 | 1187 | None |
Suzuki-Miyaura | Test 1 | 4320 | 1440 | Ligand-based |
(5760 reactions) | Test 2 | 4320 | 1440 | Ligand-based |
 | Test 3 | 4320 | 1440 | Ligand-based |
 | Test 4 | 4320 | 1440 | Ligand-based |
 | Random | 4032 | 1728 | None |