MolScribe, DECIMER HDM
Failure breakdown
Attempted: 1000 | exact: 99 (0.099)
| failure mode |
n |
share of all |
share of failures |
what it points at |
| unparseable output |
148 |
0.148 |
0.164 |
decoder emitted a string RDKit cannot read |
| stereochemistry only |
27 |
0.027 |
0.030 |
skeleton correct; wedge/hash reading wrong |
| right formula, wrong bonds |
1 |
0.001 |
0.001 |
atoms counted correctly, connectivity misread |
| other structural error |
725 |
0.725 |
0.805 |
a genuinely different molecule |
Accuracy by molecule size
| heavy atoms |
n |
exact |
valid |
| <=10 |
431 |
0.086 |
0.896 |
| <=15 |
325 |
0.123 |
0.849 |
| <=20 |
172 |
0.099 |
0.779 |
| <=25 |
70 |
0.071 |
0.771 |
| <=30 |
2 |
0.000 |
1.000 |
Sequence length
- Reference SMILES length, all attempted: median 20 tokens
- Reference SMILES length, correct predictions: median 20 tokens
Longest failures
| reference |
predicted |
C=C1C(=O)O[C@H]2/C=C(\C)CC/C=C(\C)CCC(=O)[C@@H](C)CC[C@@H]12 |
[H][C@@]12CC[C@H](C)C(=O)CC/C(C)=C/CC/C(C)=C/[C@]1(C)OC(=O)C2=C |
CSC(=S)C1=c2sc(s2)=C(C(=S)SC)SCCOCCOCCS1 |
[2H]P([13CH2]C)[13CH2]C(=O)[13NH2]C |
O=S(=O)(O)OCN(COS(=O)(=O)O)COS(=O)(=O)O |
O=[S@SP2](=O)(O)OCN(CO[S@SP1](=O)(=O)O)CO[S@SP1](=O)(=O)O |
C#C[C@@]1(O)CC2CCC3C(CCC4(C)C3CC[C@@H]4O)C2(C)C1 |
C#C[C@@]1(O)CC2CCC3C(CC[C@@]4(C)C3CCC4(C)C)C2(C)C1 |
O=C(OI(OC(=O)C(F)(F)F)C(F)(F)F)C(F)(F)F |
COC(=O)C(F)(F)F.COC(=O)C(F)(F)F.FC(F)(F)F |