{
  "id": 224104,
  "title": "Canonicalization Quirks in Train Data",
  "url": "/competitions/bms-molecular-translation/discussion/224104",
  "author_name": "",
  "post_date": "2021-03-07T01:03:48.702241200Z",
  "votes": 14,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Is there any information from the organizers about handling canonicalization in outputs? I bring this up because I've found a few examples in the dataset that aren't canonicalized.</p>\n<p>Here are some basic RDKit i/o tests:</p>\n<pre><code>from rdkit import Chem\nimport pandas as pd\nimport numpy as np\nfrom multiprocessing import Pool\nimport os\n\ndef inchi_to_mol(inchi):\n    return Chem.inchi.MolFromInchi(inchi)\n\ndef mol_to_inchi(mol):\n    return Chem.inchi.MolToInchi(mol)\n\ndf = pd.read_csv('bms-molecular-translation/train_labels.csv')\ndf = df[:500000]\n\ndef test(x):\n    return mol_to_inchi(inchi_to_mol(x))==x\n\np = Pool(processes=os.cpu_count())\nout = np.array(p.map(test, df.InChI.values))\np.close()\n\nprint(out.mean())\n&gt;0.999924\n</code></pre>\n<p>In the first 500,000 examples, there are 38 that are canonicalized differently by RDKit. The RDKit canonicalization itself is consistent. Take one of the failed examples, image ID <code>029b4b924094</code></p>\n<pre><code>inchi = df.InChI.values[24836]\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,27,29,30)'\n\n# RDKit canonicalizes differently\nmol_to_inchi(inchi_to_mol(inchi))\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == inchi\n&gt;False\n\n# RDKit canonicalization is consistent\nmol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == mol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n&gt;True\n</code></pre>\n<p>From the examples I've looked at, the canonicalization different seems to be due to the handling of charges in the InChI string.</p>\n<p>The percentage of these examples in the dataset is fairly low, but raises the question of what would happen if the RDKit-canonicalized string was submitted.</p>",
  "messages": [
    {
      "id": "1229001",
      "postDate": "03/07/2021 01:03:48",
      "content": "<p>Is there any information from the organizers about handling canonicalization in outputs? I bring this up because I've found a few examples in the dataset that aren't canonicalized.</p>\n<p>Here are some basic RDKit i/o tests:</p>\n<pre><code>from rdkit import Chem\nimport pandas as pd\nimport numpy as np\nfrom multiprocessing import Pool\nimport os\n\ndef inchi_to_mol(inchi):\n    return Chem.inchi.MolFromInchi(inchi)\n\ndef mol_to_inchi(mol):\n    return Chem.inchi.MolToInchi(mol)\n\ndf = pd.read_csv('bms-molecular-translation/train_labels.csv')\ndf = df[:500000]\n\ndef test(x):\n    return mol_to_inchi(inchi_to_mol(x))==x\n\np = Pool(processes=os.cpu_count())\nout = np.array(p.map(test, df.InChI.values))\np.close()\n\nprint(out.mean())\n&gt;0.999924\n</code></pre>\n<p>In the first 500,000 examples, there are 38 that are canonicalized differently by RDKit. The RDKit canonicalization itself is consistent. Take one of the failed examples, image ID <code>029b4b924094</code></p>\n<pre><code>inchi = df.InChI.values[24836]\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,27,29,30)'\n\n# RDKit canonicalizes differently\nmol_to_inchi(inchi_to_mol(inchi))\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == inchi\n&gt;False\n\n# RDKit canonicalization is consistent\nmol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n&gt;'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == mol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n&gt;True\n</code></pre>\n<p>From the examples I've looked at, the canonicalization different seems to be due to the handling of charges in the InChI string.</p>\n<p>The percentage of these examples in the dataset is fairly low, but raises the question of what would happen if the RDKit-canonicalized string was submitted.</p>",
      "rawMarkdown": "Is there any information from the organizers about handling canonicalization in outputs? I bring this up because I've found a few examples in the dataset that aren't canonicalized.\n\nHere are some basic RDKit i/o tests:\n\n```\nfrom rdkit import Chem\nimport pandas as pd\nimport numpy as np\nfrom multiprocessing import Pool\nimport os\n\ndef inchi_to_mol(inchi):\n    return Chem.inchi.MolFromInchi(inchi)\n\ndef mol_to_inchi(mol):\n    return Chem.inchi.MolToInchi(mol)\n\ndf = pd.read_csv('bms-molecular-translation/train_labels.csv')\ndf = df[:500000]\n\ndef test(x):\n    return mol_to_inchi(inchi_to_mol(x))==x\n\np = Pool(processes=os.cpu_count())\nout = np.array(p.map(test, df.InChI.values))\np.close()\n\nprint(out.mean())\n>0.999924\n\n```\n\nIn the first 500,000 examples, there are 38 that are canonicalized differently by RDKit. The RDKit canonicalization itself is consistent. Take one of the failed examples, image ID `029b4b924094`\n\n```\ninchi = df.InChI.values[24836]\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,27,29,30)'\n\n# RDKit canonicalizes differently\nmol_to_inchi(inchi_to_mol(inchi))\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == inchi\n>False\n\n# RDKit canonicalization is consistent\nmol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == mol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n>True\n```\n\nFrom the examples I've looked at, the canonicalization different seems to be due to the handling of charges in the InChI string.\n\nThe percentage of these examples in the dataset is fairly low, but raises the question of what would happen if the RDKit-canonicalized string was submitted.",
      "votes": null
    },
    {
      "id": "1229007",
      "postDate": "03/07/2021 01:13:29",
      "content": "<p>Great analysis and insight!</p>\n<p>To your last point of what would happen if the RDKit-canonicalized string was submitted, I imagine it would still get a fairly low Levenshtein distance, indicating you got it mostly correct. I think this issue will be largely unavoidable</p>",
      "rawMarkdown": "Great analysis and insight!\n\nTo your last point of what would happen if the RDKit-canonicalized string was submitted, I imagine it would still get a fairly low Levenshtein distance, indicating you got it mostly correct. I think this issue will be largely unavoidable",
      "votes": null
    },
    {
      "id": "1229234",
      "postDate": "03/07/2021 07:20:12",
      "content": "<p>Thank you very much for your analysis.😄<br>\nThis problem seems to be caused by the wrong InChI data. In the case of image ID 029b4b924094, the given InChI string is not consistent with the image. From the image, the chemical formula should be \"C24H25BClN6\", but the given formula is \"C24H24BClN6\". In addition, the total charge seems to be wrong. Should we remove such data in training?</p>",
      "rawMarkdown": "Thank you very much for your analysis.😄\nThis problem seems to be caused by the wrong InChI data. In the case of image ID 029b4b924094, the given InChI string is not consistent with the image. From the image, the chemical formula should be \"C24H25BClN6\", but the given formula is \"C24H24BClN6\". In addition, the total charge seems to be wrong. Should we remove such data in training?",
      "votes": null
    },
    {
      "id": "1275037",
      "postDate": "04/15/2021 20:50:16",
      "content": "<p>I found 197 of such quirks in the whole training dataset. RDkit also generates warnings explaining what's wrong. Though percentage of those formulas not so big, it may be more important as top LB score goes below 1-2.</p>\n<p>Here are 50 of them:</p>\n<table>\n<thead>\n<tr>\n<th>InChI</th>\n<th>MolToInchi(MolFromInchi(InChI))</th>\n<th>rdkit warning</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+</td>\n<td>InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+,17-3?</td>\n<td>Accepted unusual valence(s): C(3); Omitted undefined stereo</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2,15,19,20,21)</td>\n<td>InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2-3,15,16,17,18,19,20,21,22,23)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,23,24,26)</td>\n<td>InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,24,26)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C5H14BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9H,3,7-8H2,1-2H3</td>\n<td>InChI=1S/C5H15BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9-10H,3,7-8H2,1-2H3/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C17H19BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,21H,2,4,6,9-10H2,(H,20,23)</td>\n<td>InChI=1S/C17H20BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,20-21,23H,2,4,6,9-10H2/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H,26,27)</td>\n<td>InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H-2,26,27,28,29)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11H2,(H2,19,20)</td>\n<td>InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11,19H2</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8,17H,1H3,(H,14,15,18,19)</td>\n<td>InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8H,1H3,(H2,14,15,16,17)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,15,16)(H2,12,17,18)(H,13,14,19,20)</td>\n<td>InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,13,14)(H,15,16)(H2,12,17,18)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8H,1H3,(H,16,17)</td>\n<td>InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8,17H,1H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)/i15+0</td>\n<td>InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+</td>\n<td>InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+,29-23?</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15,36H,1-4H3,(H,37,38,39)</td>\n<td>InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15H,1-4H3,(H2-,33,35,36,37,38,39)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2,12,14)</td>\n<td>InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2-,12,14,15)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10H,4-6H2,1-2H3,(H,13,17)(H,14,16)(H,18,19)/t9-,10-/m0/s1</td>\n<td>InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10,17H,4-6H2,1-2H3,(H,14,16)(H,18,19)/t9-,10-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,20H,17H2,1H3,(H,18,19)</td>\n<td>InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,18,20H,17H2,1H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C6H4BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H,(H,8,9)</td>\n<td>InChI=1S/C6H3BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H/q-1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b14-13+</td>\n<td>InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b6-1?,14-13+</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1/i20+0</td>\n<td>InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H,23,27)(H,24,26)</td>\n<td>InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H2,23,24,26,27)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b14-7-/t21-,24+,25-/m1/s1</td>\n<td>InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b6-1?,14-7-/t21-,24+,25-/m1/s1</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H,19,20)</td>\n<td>InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H-,17,18,19,20)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C9H10O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3</td>\n<td>InChI=1S/C9H11O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3,(H,10,11)/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1/i10+0</td>\n<td>InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H,25,28)(H2,23,24,26,29)</td>\n<td>InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H3-,23,25,26,27,28,29)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C25H25BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17,28,33H,3,7,9-12,15H2</td>\n<td>InChI=1S/C25H24BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17H,3,7,9-12,15H2,(H-2,28,33,34,35)/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i/hD</td>\n<td>InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i7D</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C26H24BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H,29,30)(H,33,37)(H,34,38)</td>\n<td>InChI=1S/C26H23BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H2-3,29,30,31,32,33,34,37,38)/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H18BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9H,8H2,1-3H3,(H,23,24)(H,20,21,26)/b22-13-</td>\n<td>InChI=1S/C18H19BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9,25H,8H2,1-3H3,(H,23,24)(H,20,21,26)/q+1/b22-13-</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H,18,19,22)</td>\n<td>InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H-,17,18,19,22,23)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H,19,22)/t15-/m0/s1</td>\n<td>InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H-,19,22,23)/t15-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,25H,1-2H3,(H,22,23)</td>\n<td>InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,23,25H,1-2H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C23H24BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3,(H,27,34)</td>\n<td>InChI=1S/C23H23BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3/q-1/p+1</td>\n<td>Proton(s) added/removed</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14H,4,15H2,1-3H3,(H4,20,21,22,23,24)</td>\n<td>InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14,23H,4,15,20H2,1-3H3,(H,21,22)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H,7,8)(H2,4,5,6)</td>\n<td>InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H3-,3,4,5,6,7,8)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C7H14BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7H,2-6H2,1H3</td>\n<td>InChI=1S/C7H15BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7,10H,2-6H2,1H3/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i/hT</td>\n<td>InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i11T</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i/hT</td>\n<td>InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i28T</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6,12H,1H3</td>\n<td>InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6H,1H3,(H-,12,13)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H,10,11)</td>\n<td>InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H-,9,10,11)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C24H25BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18,27H,8-9,11-12,15H2,1H3</td>\n<td>InChI=1S/C24H24BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18H,8-9,11-12,15H2,1H3/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H,34,36)</td>\n<td>InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H-,34,36,37)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1/i23+0</td>\n<td>InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C19H19N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,1-2H3,(H4,20,22,23,24,25,26)</td>\n<td>InChI=1S/C19H18N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,20H2,1-2H3,(H,22,24)/q-1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H,22,23)/t15-/m0/s1</td>\n<td>InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H-,22,23,24)/t15-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,21,22,23)/t14-/m1/s1</td>\n<td>InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,22,23)/t14-/m1/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11,24H,2,7,10H2,1H3,(H2,23,25,26)</td>\n<td>InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11H,2,7,10H2,1H3,(H3-,22,23,24,25,26)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4H,1-3H3,(H,13,17)(H,15,16)</td>\n<td>InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4,17H,1-3H3,(H,15,16)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C23H22BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2,(H,30,31)</td>\n<td>InChI=1S/C23H21BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2/q-1/p+1</td>\n<td>Proton(s) added/removed; Omitted undefined stereo</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,18)</td>\n<td>InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,20)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n</tbody>\n</table>\n<p>Don't know how it is important, but RDKit is complaining about much than those 197 mismatched formulas but produces the same InChI despite of this. The most frequent warning is about undefined stereo - maybe it can be related somehow to poor performance of my network especially identifying stereolayer.</p>\n<table>\n<thead>\n<tr>\n<th>rdkit warning</th>\n<th>count   &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>Omitted undefined stereo</td>\n<td>937044</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(3); Omitted undefined stereo</td>\n<td>581</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(3)</td>\n<td>569</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3)</td>\n<td>410</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2)</td>\n<td>300</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3); Omitted undefined stereo</td>\n<td>214</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): O(1)</td>\n<td>144</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): O(1); Omitted undefined stereo</td>\n<td>121</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); Omitted undefined stereo</td>\n<td>116</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1)</td>\n<td>112</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1); Omitted undefined stereo</td>\n<td>112</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(2)</td>\n<td>67</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(2)</td>\n<td>63</td>\n</tr>\n<tr>\n<td>Charges were rearranged</td>\n<td>54</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(2); Omitted undefined stereo</td>\n<td>39</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n<td>39</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(2); Omitted undefined stereo</td>\n<td>37</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(2); Omitted undefined stereo</td>\n<td>35</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(2)</td>\n<td>28</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Omitted undefined stereo</td>\n<td>27</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n<td>24</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n<td>20</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2)</td>\n<td>14</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n<td>13</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(3)</td>\n<td>11</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)</td>\n<td>10</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n<td>8</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1); Omitted undefined stereo</td>\n<td>6</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(3); Omitted undefined stereo</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): O(4); Omitted undefined stereo</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(1)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2); Omitted undefined stereo</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): O(4)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Proton(s) added/removed</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); C+1(2)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(1)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); C+1(2); Omitted undefined stereo</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1); B+1(1); Omitted undefined stereo</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number; Proton balance: Does not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): I(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si(3)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Exchangeable isotopic H: Do not match; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); B+1(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): C-1(2); C(3); O(4)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): P(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3); B+1(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Proton(s) added/removed; Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); B+1(1)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1); O(1)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): C(3); O(4)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si-1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); Si+1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C+1(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C+1(2); C(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2); Si(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Multiple instead of one, Attachment points, Number; Charge(s): Do not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si(3)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Stereobonds/cumulenes: Extra undefined)</td>\n<td>1</td>\n</tr>\n</tbody>\n</table>",
      "rawMarkdown": "I found 197 of such quirks in the whole training dataset. RDkit also generates warnings explaining what's wrong. Though percentage of those formulas not so big, it may be more important as top LB score goes below 1-2.\n\nHere are 50 of them:\n| InChI                                                                                                                                                   | MolToInchi(MolFromInchi(InChI))                                                                                                                                                                                              | rdkit warning                                                                                                                                                                    |\n|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+                                                                          | InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+,17-3?                                                                             | Accepted unusual valence(s): C(3); Omitted undefined stereo                                                                                                                          |\n| InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2,15,19,20,21)                                                                                      | InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2-3,15,16,17,18,19,20,21,22,23)                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,23,24,26)                                                             | InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,24,26)                                                                         | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C5H14BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9H,3,7-8H2,1-2H3                                                                                                                                        | InChI=1S/C5H15BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9-10H,3,7-8H2,1-2H3/q+1                                                                                                                                          | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C17H19BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,21H,2,4,6,9-10H2,(H,20,23)                                                              | InChI=1S/C17H20BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,20-21,23H,2,4,6,9-10H2/q+1                                                                       | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                         |\n| InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H,26,27)                                                 | InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H-2,26,27,28,29)                                                  | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11H2,(H2,19,20)                                                                                   | InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11,19H2                                                                                                    | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8,17H,1H3,(H,14,15,18,19)                                                                                                 | InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8H,1H3,(H2,14,15,16,17)                                                                                                            | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,15,16)(H2,12,17,18)(H,13,14,19,20)                                                                     | InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,13,14)(H,15,16)(H2,12,17,18)                                                                                    | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8H,1H3,(H,16,17)                                                                                                              | InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8,17H,1H3                                                                                                                              | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)/i15+0                                                | InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)                                                               |                                                                                                                                                                                      |\n| InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+                                           | InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+,29-23?                                             | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15,36H,1-4H3,(H,37,38,39)                        | InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15H,1-4H3,(H2-,33,35,36,37,38,39)                         | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2,12,14)                                                                                                                      | InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2-,12,14,15)                                                                                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10H,4-6H2,1-2H3,(H,13,17)(H,14,16)(H,18,19)/t9-,10-/m0/s1                                                              | InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10,17H,4-6H2,1-2H3,(H,14,16)(H,18,19)/t9-,10-/m0/s1                                                                             | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,20H,17H2,1H3,(H,18,19)                                                                                            | InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,18,20H,17H2,1H3                                                                                                            | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C6H4BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H,(H,8,9)                                                                                                                                              | InChI=1S/C6H3BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H/q-1                                                                                                                                                           | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                              |\n| InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b14-13+                                                                                                     | InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b6-1?,14-13+                                                                                                         | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1/i20+0                                                            | InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1                                                                           |                                                                                                                                                                                      |\n| InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H,23,27)(H,24,26)                                                | InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H2,23,24,26,27)                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)                                                                                                     |\n| InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b14-7-/t21-,24+,25-/m1/s1                                    | InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b6-1?,14-7-/t21-,24+,25-/m1/s1                                        | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H,19,20)                                                                                                            | InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H-,17,18,19,20)                                                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C9H10O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3                                                                                                                                           | InChI=1S/C9H11O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3,(H,10,11)/q+1                                                                                                                                      | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                                      |\n| InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1/i10+0                                                                                                   | InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1                                                                                                                  |                                                                                                                                                                                      |\n| InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H,25,28)(H2,23,24,26,29)                                                 | InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H3-,23,25,26,27,28,29)                                                            | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                  |\n| InChI=1S/C25H25BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17,28,33H,3,7,9-12,15H2                           | InChI=1S/C25H24BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17H,3,7,9-12,15H2,(H-2,28,33,34,35)/q-1/p+1                | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match) |\n| InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i/hD                                                       | InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i7D                                                                 | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C26H24BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H,29,30)(H,33,37)(H,34,38)   | InChI=1S/C26H23BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H2-3,29,30,31,32,33,34,37,38)/q-1/p+1 | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)      |\n| InChI=1S/C18H18BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9H,8H2,1-3H3,(H,23,24)(H,20,21,26)/b22-13-                                               | InChI=1S/C18H19BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9,25H,8H2,1-3H3,(H,23,24)(H,20,21,26)/q+1/b22-13-                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H,18,19,22)                                                                               | InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H-,17,18,19,22,23)                                                                                 | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H,19,22)/t15-/m0/s1                                                             | InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H-,19,22,23)/t15-/m0/s1                                                                  | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,25H,1-2H3,(H,22,23)                                                                                  | InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,23,25H,1-2H3                                                                                                  | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C23H24BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3,(H,27,34)                                | InChI=1S/C23H23BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3/q-1/p+1                                           | Proton(s) added/removed                                                                                                                                                              |\n| InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14H,4,15H2,1-3H3,(H4,20,21,22,23,24)                                                                    | InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14,23H,4,15,20H2,1-3H3,(H,21,22)                                                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H,7,8)(H2,4,5,6)                                                                                                                                     | InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H3-,3,4,5,6,7,8)                                                                                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                  |\n| InChI=1S/C7H14BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7H,2-6H2,1H3                                                                                                                                             | InChI=1S/C7H15BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7,10H,2-6H2,1H3/q+1                                                                                                                                               | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i/hT                                                                                                                                    | InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i11T                                                                                                                                             | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i/hT                                           | InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i28T                                                    | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6,12H,1H3                                                                                                                                | InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6H,1H3,(H-,12,13)                                                                                                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)                                                                  |\n| InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H,10,11)                                                                                                                                           | InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H-,9,10,11)                                                                                                                                                 | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C24H25BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18,27H,8-9,11-12,15H2,1H3                     | InChI=1S/C24H24BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18H,8-9,11-12,15H2,1H3/q-1/p+1                         | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)                                                      |\n| InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H,34,36)              | InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H-,34,36,37)                   | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1/i23+0 | InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1                |                                                                                                                                                                                      |\n| InChI=1S/C19H19N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,1-2H3,(H4,20,22,23,24,25,26)                                                      | InChI=1S/C19H18N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,20H2,1-2H3,(H,22,24)/q-1                                                                   | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)                                          |\n| InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H,22,23)/t15-/m0/s1                                                        | InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H-,22,23,24)/t15-/m0/s1                                                             | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,21,22,23)/t14-/m1/s1                                                           | InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,22,23)/t14-/m1/s1                                                                       | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11,24H,2,7,10H2,1H3,(H2,23,25,26)                                                           | InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11H,2,7,10H2,1H3,(H3-,22,23,24,25,26)                                                                | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4H,1-3H3,(H,13,17)(H,15,16)                                                                                                            | InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4,17H,1-3H3,(H,15,16)                                                                                                                           | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)                                                  |\n| InChI=1S/C23H22BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2,(H,30,31)                                                     | InChI=1S/C23H21BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2/q-1/p+1                                                                | Proton(s) added/removed; Omitted undefined stereo                                                                                                                                    |\n| InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,18)                                                                                                  | InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,20)                                                                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n\nDon't know how it is important, but RDKit is complaining about much than those 197 mismatched formulas but produces the same InChI despite of this. The most frequent warning is about undefined stereo - maybe it can be related somehow to poor performance of my network especially identifying stereolayer.\n\n| rdkit warning                                                                                                                                                                                          |   count   &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |\n|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------:|\n| Omitted undefined stereo                                                                                                                                                                       |              937044 |\n| Accepted unusual valence(s): Si(3); Omitted undefined stereo                                                                                                                                   |                 581 |\n| Accepted unusual valence(s): Si(3)                                                                                                                                                             |                 569 |\n| Accepted unusual valence(s): C(3)                                                                                                                                                              |                 410 |\n| Accepted unusual valence(s): B(2)                                                                                                                                                              |                 300 |\n| Accepted unusual valence(s): C(3); Omitted undefined stereo                                                                                                                                    |                 214 |\n| Accepted unusual valence(s): O(1)                                                                                                                                                              |                 144 |\n| Accepted unusual valence(s): O(1); Omitted undefined stereo                                                                                                                                    |                 121 |\n| Accepted unusual valence(s): B(2); Omitted undefined stereo                                                                                                                                    |                 116 |\n| Accepted unusual valence(s): B(1)                                                                                                                                                              |                 112 |\n| Accepted unusual valence(s): B(1); Omitted undefined stereo                                                                                                                                    |                 112 |\n| Accepted unusual valence(s): N(2)                                                                                                                                                              |                  67 |\n| Accepted unusual valence(s): Si(2)                                                                                                                                                             |                  63 |\n| Charges were rearranged                                                                                                                                                                        |                  54 |\n| Accepted unusual valence(s): N(2); Omitted undefined stereo                                                                                                                                    |                  39 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                                     |                  39 |\n| Accepted unusual valence(s): Si(2); Omitted undefined stereo                                                                                                                                   |                  37 |\n| Accepted unusual valence(s): C(2); Omitted undefined stereo                                                                                                                                    |                  35 |\n| Accepted unusual valence(s): C(2)                                                                                                                                                              |                  28 |\n| Charges were rearranged; Omitted undefined stereo                                                                                                                                              |                  27 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                                    |                  24 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                                |                  20 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2)                                                                                                                                                  |                  14 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                                     |                  13 |\n| Accepted unusual valence(s): S(3)                                                                                                                                                              |                  11 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)                                                                            |                  10 |\n| Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                            |                   8 |\n| Accepted unusual valence(s): B+1(1); Omitted undefined stereo                                                                                                                                  |                   6 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                            |                   5 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)                                                            |                   5 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                                        |                   5 |\n| Accepted unusual valence(s): S(3); Omitted undefined stereo                                                                                                                                    |                   4 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                   |                   4 |\n| Charges were rearranged; Accepted unusual valence(s): O(4); Omitted undefined stereo                                                                                                           |                   4 |\n| Accepted unusual valence(s): B+1(1)                                                                                                                                                            |                   4 |\n| Accepted unusual valence(s): S(1)                                                                                                                                                              |                   4 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2); Omitted undefined stereo                                                                                                                        |                   3 |\n| Charges were rearranged; Accepted unusual valence(s): O(4)                                                                                                                                     |                   3 |\n| Proton(s) added/removed                                                                                                                                                                        |                   3 |\n| Accepted unusual valence(s): C-1(2); C+1(2)                                                                                                                                                    |                   3 |\n| Accepted unusual valence(s): N(1)                                                                                                                                                              |                   3 |\n| Accepted unusual valence(s): C-1(2); C+1(2); Omitted undefined stereo                                                                                                                          |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                                                |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)                                                    |                   2 |\n| Accepted unusual valence(s): B(1); B+1(1); Omitted undefined stereo                                                                                                                            |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)                                                                |                   2 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number; Proton balance: Does not match)                                                                                     |                   2 |\n| Accepted unusual valence(s): N(1); Omitted undefined stereo                                                                                                                                    |                   1 |\n| Accepted unusual valence(s): I(2)                                                                                                                                                              |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points; Proton balance: Does not match)                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number; Proton balance: Does not match)                                                     |                   1 |\n| Accepted unusual valence(s): Si-1(2)                                                                                                                                                           |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si(3)                                                                                                                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Number; Charge(s): Do not match; Proton balance: Does not match)                                   |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number)                                                                                                                                 |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match; Proton balance: Does not match)                                        |                   1 |\n| Accepted unusual valence(s): Si-1(2); Omitted undefined stereo                                                                                                                                 |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Exchangeable isotopic H: Do not match; Charge(s): Do not match; Proton balance: Does not match) |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)                |                   1 |\n| Accepted unusual valence(s): B(2); B+1(1); Omitted undefined stereo                                                                                                                            |                   1 |\n| Charges were rearranged; Accepted unusual valence(s): C-1(2); C(3); O(4)                                                                                                                       |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si(2); Omitted undefined stereo                                                                                                                          |                   1 |\n| Accepted unusual valence(s): P(2)                                                                                                                                                              |                   1 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)                                                                                                               |                   1 |\n| Accepted unusual valence(s): C(3); B+1(1); Omitted undefined stereo                                                                                                                            |                   1 |\n| Proton(s) added/removed; Omitted undefined stereo                                                                                                                                              |                   1 |\n| Accepted unusual valence(s): B(2); B+1(1)                                                                                                                                                      |                   1 |\n| Accepted unusual valence(s): B+1(1); O(1)                                                                                                                                                      |                   1 |\n| Charges were rearranged; Accepted unusual valence(s): C(3); O(4)                                                                                                                               |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si-1(2)                                                                                                                                                  |                   1 |\n| Accepted unusual valence(s): C-1(2); Si+1(2)                                                                                                                                                   |                   1 |\n| Accepted unusual valence(s): C+1(2); Omitted undefined stereo                                                                                                                                  |                   1 |\n| Accepted unusual valence(s): C+1(2); C(2); Omitted undefined stereo                                                                                                                            |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2); Si(2)                                                                                                                                           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Multiple instead of one, Attachment points, Number; Charge(s): Do not match)                                                |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si(3)                                                                                                                                                    |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si(2)                                                                                                                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Stereobonds/cumulenes: Extra undefined)                                                                                                                         |                   1 |",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1229007,
      "author_name": "matthewmasters",
      "author_url": "",
      "post_date": "03/07/2021 01:13:29",
      "content": "<p>Great analysis and insight!</p>\n<p>To your last point of what would happen if the RDKit-canonicalized string was submitted, I imagine it would still get a fairly low Levenshtein distance, indicating you got it mostly correct. I think this issue will be largely unavoidable</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1229234,
      "author_name": "hiroshisakiyama",
      "author_url": "",
      "post_date": "03/07/2021 07:20:12",
      "content": "<p>Thank you very much for your analysis.😄<br>\nThis problem seems to be caused by the wrong InChI data. In the case of image ID 029b4b924094, the given InChI string is not consistent with the image. From the image, the chemical formula should be \"C24H25BClN6\", but the given formula is \"C24H24BClN6\". In addition, the total charge seems to be wrong. Should we remove such data in training?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1275037,
      "author_name": "stassl",
      "author_url": "",
      "post_date": "04/15/2021 20:50:16",
      "content": "<p>I found 197 of such quirks in the whole training dataset. RDkit also generates warnings explaining what's wrong. Though percentage of those formulas not so big, it may be more important as top LB score goes below 1-2.</p>\n<p>Here are 50 of them:</p>\n<table>\n<thead>\n<tr>\n<th>InChI</th>\n<th>MolToInchi(MolFromInchi(InChI))</th>\n<th>rdkit warning</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+</td>\n<td>InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+,17-3?</td>\n<td>Accepted unusual valence(s): C(3); Omitted undefined stereo</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2,15,19,20,21)</td>\n<td>InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2-3,15,16,17,18,19,20,21,22,23)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,23,24,26)</td>\n<td>InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,24,26)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C5H14BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9H,3,7-8H2,1-2H3</td>\n<td>InChI=1S/C5H15BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9-10H,3,7-8H2,1-2H3/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C17H19BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,21H,2,4,6,9-10H2,(H,20,23)</td>\n<td>InChI=1S/C17H20BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,20-21,23H,2,4,6,9-10H2/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H,26,27)</td>\n<td>InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H-2,26,27,28,29)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11H2,(H2,19,20)</td>\n<td>InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11,19H2</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8,17H,1H3,(H,14,15,18,19)</td>\n<td>InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8H,1H3,(H2,14,15,16,17)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,15,16)(H2,12,17,18)(H,13,14,19,20)</td>\n<td>InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,13,14)(H,15,16)(H2,12,17,18)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8H,1H3,(H,16,17)</td>\n<td>InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8,17H,1H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)/i15+0</td>\n<td>InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+</td>\n<td>InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+,29-23?</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15,36H,1-4H3,(H,37,38,39)</td>\n<td>InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15H,1-4H3,(H2-,33,35,36,37,38,39)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2,12,14)</td>\n<td>InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2-,12,14,15)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10H,4-6H2,1-2H3,(H,13,17)(H,14,16)(H,18,19)/t9-,10-/m0/s1</td>\n<td>InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10,17H,4-6H2,1-2H3,(H,14,16)(H,18,19)/t9-,10-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,20H,17H2,1H3,(H,18,19)</td>\n<td>InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,18,20H,17H2,1H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C6H4BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H,(H,8,9)</td>\n<td>InChI=1S/C6H3BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H/q-1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b14-13+</td>\n<td>InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b6-1?,14-13+</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1/i20+0</td>\n<td>InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H,23,27)(H,24,26)</td>\n<td>InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H2,23,24,26,27)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b14-7-/t21-,24+,25-/m1/s1</td>\n<td>InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b6-1?,14-7-/t21-,24+,25-/m1/s1</td>\n<td>Accepted unusual valence(s): C(3)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H,19,20)</td>\n<td>InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H-,17,18,19,20)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C9H10O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3</td>\n<td>InChI=1S/C9H11O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3,(H,10,11)/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1/i10+0</td>\n<td>InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H,25,28)(H2,23,24,26,29)</td>\n<td>InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H3-,23,25,26,27,28,29)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C25H25BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17,28,33H,3,7,9-12,15H2</td>\n<td>InChI=1S/C25H24BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17H,3,7,9-12,15H2,(H-2,28,33,34,35)/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i/hD</td>\n<td>InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i7D</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C26H24BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H,29,30)(H,33,37)(H,34,38)</td>\n<td>InChI=1S/C26H23BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H2-3,29,30,31,32,33,34,37,38)/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H18BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9H,8H2,1-3H3,(H,23,24)(H,20,21,26)/b22-13-</td>\n<td>InChI=1S/C18H19BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9,25H,8H2,1-3H3,(H,23,24)(H,20,21,26)/q+1/b22-13-</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H,18,19,22)</td>\n<td>InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H-,17,18,19,22,23)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H,19,22)/t15-/m0/s1</td>\n<td>InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H-,19,22,23)/t15-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,25H,1-2H3,(H,22,23)</td>\n<td>InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,23,25H,1-2H3</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C23H24BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3,(H,27,34)</td>\n<td>InChI=1S/C23H23BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3/q-1/p+1</td>\n<td>Proton(s) added/removed</td>\n</tr>\n<tr>\n<td>InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14H,4,15H2,1-3H3,(H4,20,21,22,23,24)</td>\n<td>InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14,23H,4,15,20H2,1-3H3,(H,21,22)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H,7,8)(H2,4,5,6)</td>\n<td>InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H3-,3,4,5,6,7,8)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C7H14BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7H,2-6H2,1H3</td>\n<td>InChI=1S/C7H15BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7,10H,2-6H2,1H3/q+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i/hT</td>\n<td>InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i11T</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i/hT</td>\n<td>InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i28T</td>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6,12H,1H3</td>\n<td>InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6H,1H3,(H-,12,13)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)</td>\n</tr>\n<tr>\n<td>InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H,10,11)</td>\n<td>InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H-,9,10,11)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C24H25BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18,27H,8-9,11-12,15H2,1H3</td>\n<td>InChI=1S/C24H24BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18H,8-9,11-12,15H2,1H3/q-1/p+1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H,34,36)</td>\n<td>InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H-,34,36,37)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1/i23+0</td>\n<td>InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1</td>\n<td></td>\n</tr>\n<tr>\n<td>InChI=1S/C19H19N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,1-2H3,(H4,20,22,23,24,25,26)</td>\n<td>InChI=1S/C19H18N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,20H2,1-2H3,(H,22,24)/q-1</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H,22,23)/t15-/m0/s1</td>\n<td>InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H-,22,23,24)/t15-/m0/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,21,22,23)/t14-/m1/s1</td>\n<td>InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,22,23)/t14-/m1/s1</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11,24H,2,7,10H2,1H3,(H2,23,25,26)</td>\n<td>InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11H,2,7,10H2,1H3,(H3-,22,23,24,25,26)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4H,1-3H3,(H,13,17)(H,15,16)</td>\n<td>InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4,17H,1-3H3,(H,15,16)</td>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n</tr>\n<tr>\n<td>InChI=1S/C23H22BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2,(H,30,31)</td>\n<td>InChI=1S/C23H21BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2/q-1/p+1</td>\n<td>Proton(s) added/removed; Omitted undefined stereo</td>\n</tr>\n<tr>\n<td>InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,18)</td>\n<td>InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,20)</td>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n</tr>\n</tbody>\n</table>\n<p>Don't know how it is important, but RDKit is complaining about much than those 197 mismatched formulas but produces the same InChI despite of this. The most frequent warning is about undefined stereo - maybe it can be related somehow to poor performance of my network especially identifying stereolayer.</p>\n<table>\n<thead>\n<tr>\n<th>rdkit warning</th>\n<th>count   &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>Omitted undefined stereo</td>\n<td>937044</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(3); Omitted undefined stereo</td>\n<td>581</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(3)</td>\n<td>569</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3)</td>\n<td>410</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2)</td>\n<td>300</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3); Omitted undefined stereo</td>\n<td>214</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): O(1)</td>\n<td>144</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): O(1); Omitted undefined stereo</td>\n<td>121</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); Omitted undefined stereo</td>\n<td>116</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1)</td>\n<td>112</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1); Omitted undefined stereo</td>\n<td>112</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(2)</td>\n<td>67</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(2)</td>\n<td>63</td>\n</tr>\n<tr>\n<td>Charges were rearranged</td>\n<td>54</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(2); Omitted undefined stereo</td>\n<td>39</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)</td>\n<td>39</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si(2); Omitted undefined stereo</td>\n<td>37</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(2); Omitted undefined stereo</td>\n<td>35</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(2)</td>\n<td>28</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Omitted undefined stereo</td>\n<td>27</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)</td>\n<td>24</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)</td>\n<td>20</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2)</td>\n<td>14</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)</td>\n<td>13</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(3)</td>\n<td>11</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)</td>\n<td>10</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)</td>\n<td>8</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1); Omitted undefined stereo</td>\n<td>6</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n<td>5</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(3); Omitted undefined stereo</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): O(4); Omitted undefined stereo</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): S(1)</td>\n<td>4</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2); Omitted undefined stereo</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): O(4)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Proton(s) added/removed</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); C+1(2)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(1)</td>\n<td>3</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); C+1(2); Omitted undefined stereo</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(1); B+1(1); Omitted undefined stereo</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number; Proton balance: Does not match)</td>\n<td>2</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): N(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): I(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si(3)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Exchangeable isotopic H: Do not match; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); B+1(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): C-1(2); C(3); O(4)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): P(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C(3); B+1(1); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Proton(s) added/removed; Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B(2); B+1(1)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): B+1(1); O(1)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Charges were rearranged; Accepted unusual valence(s): C(3); O(4)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si-1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C-1(2); Si+1(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C+1(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): C+1(2); C(2); Omitted undefined stereo</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si-1(2); Si+1(2); Si(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Multiple instead of one, Attachment points, Number; Charge(s): Do not match)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si(3)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Accepted unusual valence(s): Si+1(2); Si(2)</td>\n<td>1</td>\n</tr>\n<tr>\n<td>Problems/mismatches: Mobile-H( Stereobonds/cumulenes: Extra undefined)</td>\n<td>1</td>\n</tr>\n</tbody>\n</table>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1229001": "Is there any information from the organizers about handling canonicalization in outputs? I bring this up because I've found a few examples in the dataset that aren't canonicalized.\n\nHere are some basic RDKit i/o tests:\n\n```\nfrom rdkit import Chem\nimport pandas as pd\nimport numpy as np\nfrom multiprocessing import Pool\nimport os\n\ndef inchi_to_mol(inchi):\n    return Chem.inchi.MolFromInchi(inchi)\n\ndef mol_to_inchi(mol):\n    return Chem.inchi.MolToInchi(mol)\n\ndf = pd.read_csv('bms-molecular-translation/train_labels.csv')\ndf = df[:500000]\n\ndef test(x):\n    return mol_to_inchi(inchi_to_mol(x))==x\n\np = Pool(processes=os.cpu_count())\nout = np.array(p.map(test, df.InChI.values))\np.close()\n\nprint(out.mean())\n>0.999924\n\n```\n\nIn the first 500,000 examples, there are 38 that are canonicalized differently by RDKit. The RDKit canonicalization itself is consistent. Take one of the failed examples, image ID `029b4b924094`\n\n```\ninchi = df.InChI.values[24836]\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,27,29,30)'\n\n# RDKit canonicalizes differently\nmol_to_inchi(inchi_to_mol(inchi))\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == inchi\n>False\n\n# RDKit canonicalization is consistent\nmol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n>'InChI=1S/C24H24BClN6/c1-31-15-19(13-28-31)17-7-16(14-32-5-3-2-4-6-32)8-21(10-17)29-24-27-12-18-9-20(25)11-22(26)23(18)30-24/h7-13,15H,2-6,14H2,1H3,(H,29,30)/p+1'\n\nmol_to_inchi(inchi_to_mol(inchi)) == mol_to_inchi(inchi_to_mol(mol_to_inchi(inchi_to_mol(inchi))))\n>True\n```\n\nFrom the examples I've looked at, the canonicalization different seems to be due to the handling of charges in the InChI string.\n\nThe percentage of these examples in the dataset is fairly low, but raises the question of what would happen if the RDKit-canonicalized string was submitted.",
    "1229007": "Great analysis and insight!\n\nTo your last point of what would happen if the RDKit-canonicalized string was submitted, I imagine it would still get a fairly low Levenshtein distance, indicating you got it mostly correct. I think this issue will be largely unavoidable",
    "1229234": "Thank you very much for your analysis.😄\nThis problem seems to be caused by the wrong InChI data. In the case of image ID 029b4b924094, the given InChI string is not consistent with the image. From the image, the chemical formula should be \"C24H25BClN6\", but the given formula is \"C24H24BClN6\". In addition, the total charge seems to be wrong. Should we remove such data in training?",
    "1275037": "I found 197 of such quirks in the whole training dataset. RDkit also generates warnings explaining what's wrong. Though percentage of those formulas not so big, it may be more important as top LB score goes below 1-2.\n\nHere are 50 of them:\n| InChI                                                                                                                                                   | MolToInchi(MolFromInchi(InChI))                                                                                                                                                                                              | rdkit warning                                                                                                                                                                    |\n|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+                                                                          | InChI=1S/C21H43O2Si2/c1-13-25(14-2,18(5)6)22-19(7)15-16-20(17(3)4)23-24(11,12)21(8,9)10/h3,15-16,18-20H,13-14H2,1-2,4-12H3/b16-15+,17-3?                                                                             | Accepted unusual valence(s): C(3); Omitted undefined stereo                                                                                                                          |\n| InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2,15,19,20,21)                                                                                      | InChI=1S/C14H13N9O/c1-8-17-12(22-13-20-10(15)19-11(16-2)21-13)23-14(18-8)24-9-6-4-3-5-7-9/h3-7H,1-2H3,(H2-3,15,16,17,18,19,20,21,22,23)                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,23,24,26)                                                             | InChI=1S/C16H17BF3N5OS/c1-2-21-15(26)24-12-5-9(10(17)6-23-12)14-25-13(16(18,19)20)11(27-14)7-22-8-3-4-8/h5-6,8,22H,2-4,7H2,1H3,(H2,21,24,26)                                                                         | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C5H14BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9H,3,7-8H2,1-2H3                                                                                                                                        | InChI=1S/C5H15BN3O2S/c1-3-4(7)5(6,8)12(10,11)9-2/h4,9-10H,3,7-8H2,1-2H3/q+1                                                                                                                                          | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C17H19BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,21H,2,4,6,9-10H2,(H,20,23)                                                              | InChI=1S/C17H20BN6O/c18-14-11-22-24-16(21-9-12-3-1-5-19-8-12)7-15(23-17(14)24)20-10-13-4-2-6-25-13/h1,3,5,7-8,11,13,20-21,23H,2,4,6,9-10H2/q+1                                                                       | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                         |\n| InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H,26,27)                                                 | InChI=1S/C19H24BN5O4S/c1-3-25(18-21-12-15(20)13-22-18)10-9-24(2)30(28,29)23-19(17(26)27)11-16(19)14-7-5-4-6-8-14/h4-8,12-13,16,23H,3,9-11H2,1-2H3,(H-2,26,27,28,29)                                                  | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11H2,(H2,19,20)                                                                                   | InChI=1S/C18H22N2/c1-13-12-16(19)18-15(8-5-9-17(18)20-13)11-10-14-6-3-2-4-7-14/h1,5,8-9,12,14H,2-4,6-7,10-11,19H2                                                                                                    | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8,17H,1H3,(H,14,15,18,19)                                                                                                 | InChI=1S/C13H12N2O4S/c1-9(16)10-2-7-13(14-8-10)15-20(18,19)12-5-3-11(17)4-6-12/h2-8H,1H3,(H2,14,15,16,17)                                                                                                            | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,15,16)(H2,12,17,18)(H,13,14,19,20)                                                                     | InChI=1S/C10H10ClN3O6S2/c11-5-3-6-8(4-7(5)21(12,17)18)22(19,20)14-9(13-6)1-2-10(15)16/h3-4H,1-2H2,(H,13,14)(H,15,16)(H2,12,17,18)                                                                                    | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8H,1H3,(H,16,17)                                                                                                              | InChI=1S/C14H11BN2O/c1-18-13-5-3-2-4-10(13)12-8-17-14-11(12)6-9(15)7-16-14/h2-8,17H,1H3                                                                                                                              | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)/i15+0                                                | InChI=1S/C14H11IN2O9P2/c15-7-1-4-11-10(5-7)14(18)17-13(16-11)9-3-2-8(25-27(19,20)21)6-12(9)26-28(22,23)24/h1-6H,(H,16,17,18)(H2,19,20,21)(H2,22,23,24)                                                               |                                                                                                                                                                                      |\n| InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+                                           | InChI=1S/C29H48BO/c1-4-6-16-26(17-7-5-2)18-14-23-29(24-15-25-31-3)30(27-19-10-8-11-20-27)28-21-12-9-13-22-28/h14-15,24,27-28H,4-13,16-17,19-22,25H2,1-3H3/b24-15+,29-23?                                             | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15,36H,1-4H3,(H,37,38,39)                        | InChI=1S/C30H25BN4O5S/c1-16-11-17(2)27(26(12-16)41(37,38)39)33-32-24-15-25(40-4)29(23-8-6-5-7-22(23)24)35-34-28-18(3)13-19-14-20(31)9-10-21(19)30(28)36/h5-15H,1-4H3,(H2-,33,35,36,37,38,39)                         | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2,12,14)                                                                                                                      | InChI=1S/C10H9BN2O2S/c1-5(9(12)14)13-7-3-2-6(11)4-8(7)16-10(13)15/h2-5H,1H3,(H2-,12,14,15)                                                                                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10H,4-6H2,1-2H3,(H,13,17)(H,14,16)(H,18,19)/t9-,10-/m0/s1                                                              | InChI=1S/C12H18N2O5S3/c1-7(15)4-9(12(18)19)5-20-22-21-6-10(11(17)13-3)14-8(2)16/h3,9-10,17H,4-6H2,1-2H3,(H,14,16)(H,18,19)/t9-,10-/m0/s1                                                                             | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,20H,17H2,1H3,(H,18,19)                                                                                            | InChI=1S/C15H16BN3O2/c1-21-15(19-16)18-14(10-5-7-13(20)8-6-10)11-3-2-4-12(17)9-11/h2-9,14,18,20H,17H2,1H3                                                                                                            | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C6H4BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H,(H,8,9)                                                                                                                                              | InChI=1S/C6H3BrNOS2/c7-4-1-2-5-6(3-4)10-11(9)8-5/h1-3H/q-1                                                                                                                                                           | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                              |\n| InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b14-13+                                                                                                     | InChI=1S/C16H28BO2/c1-6-7-8-9-10-11-12-13-14-17-18-15(2,3)16(4,5)19-17/h1,6,13-14H,7-12H2,2-5H3/b6-1?,14-13+                                                                                                         | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1/i20+0                                                            | InChI=1S/C19H26FNO3/c1-23-18-9-13-6-8-21-12-14(5-3-4-7-20)17(22)11-16(21)15(13)10-19(18)24-2/h9-10,14,16H,3-8,11-12H2,1-2H3/t14-,16-/m1/s1                                                                           |                                                                                                                                                                                      |\n| InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H,23,27)(H,24,26)                                                | InChI=1S/C20H32BN5O5/c1-17(2)25-26-19-6-5-18(15-24-19)20(27)23-8-4-10-29-12-14-30-13-11-28-9-3-7-22-16-31-21/h5-6,15-16H,3-4,7-14H2,1-2H3,(H2,23,24,26,27)                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)                                                                                                     |\n| InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b14-7-/t21-,24+,25-/m1/s1                                    | InChI=1S/C26H33O3Si/c1-6-7-14-21-19-25(27-5)29-24(21)20-28-30(26(2,3)4,22-15-10-8-11-16-22)23-17-12-9-13-18-23/h1,6-18,21,24-25H,19-20H2,2-5H3/b6-1?,14-7-/t21-,24+,25-/m1/s1                                        | Accepted unusual valence(s): C(3)                                                                                                                                                    |\n| InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H,19,20)                                                                                                            | InChI=1S/C15H9BN2O2/c16-10-5-3-9(4-6-10)13-8-12(15(19)20)11-2-1-7-17-14(11)18-13/h1-8H,(H-,17,18,19,20)                                                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C9H10O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3                                                                                                                                           | InChI=1S/C9H11O4S/c1-7-4-5-8(12-2)9(6-7)14(10,11)13-3/h3-6H,1-2H3,(H,10,11)/q+1                                                                                                                                      | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                                      |\n| InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1/i10+0                                                                                                   | InChI=1S/C11H16O5/c1-11(2)15-8-5-6(10(13)14-3)4-7(12)9(8)16-11/h5,7-9,12H,4H2,1-3H3/t7-,8-,9+/m1/s1                                                                                                                  |                                                                                                                                                                                      |\n| InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H,25,28)(H2,23,24,26,29)                                                 | InChI=1S/C18H19BF3N5O2S/c1-2-23-17(29)26-12-7-10(11(19)8-24-12)16-27-14(18(20,21)22)13(30-16)15(28)25-9-5-3-4-6-9/h7-9H,2-6H2,1H3,(H3-,23,25,26,27,28,29)                                                            | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                  |\n| InChI=1S/C25H25BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17,28,33H,3,7,9-12,15H2                           | InChI=1S/C25H24BN6O3S/c26-20-16-29-32-24(13-21(30-25(20)32)19-6-2-3-7-22(19)33)28-15-17-9-11-31(12-10-17)36(34,35)23-8-4-1-5-18(23)14-27/h1-2,4-6,8,13,16-17H,3,7,9-12,15H2,(H-2,28,33,34,35)/q-1/p+1                | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match) |\n| InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i/hD                                                       | InChI=1S/C6H14O13P2/c7-3(1-18-20(12,13)14)4(8)6(11,5(9)10)2-19-21(15,16)17/h3-4,7-8,11H,1-2H2,(H,9,10)(H2,12,13,14)(H2,15,16,17)/t3-,4-,6+/m1/s1/i7D                                                                 | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C26H24BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H,29,30)(H,33,37)(H,34,38)   | InChI=1S/C26H23BFN8O3S/c27-15-2-4-18(17(28)12-15)36-23-20-21(34-26(36)38)22(40-25(20)32-14-31-23)24(37)33-16-3-5-19(30-13-16)29-6-1-7-35-8-10-39-11-9-35/h2-5,12-14H,1,6-11H2,(H2-3,29,30,31,32,33,34,37,38)/q-1/p+1 | Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)      |\n| InChI=1S/C18H18BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9H,8H2,1-3H3,(H,23,24)(H,20,21,26)/b22-13-                                               | InChI=1S/C18H19BN3O7S/c1-18(2,3)28-17(26)21-16-20-12(9-30-16)13(14(23)24)22-27-8-10-5-4-6-11(7-10)15(25)29-19/h4-7,9,25H,8H2,1-3H3,(H,23,24)(H,20,21,26)/q+1/b22-13-                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H,18,19,22)                                                                               | InChI=1S/C15H18BN5O2/c16-15(23)20-7-9(8-20)12-18-13-11(14(22)19-12)6-17-21(13)10-4-2-1-3-5-10/h6,9-10H,1-5,7-8H2,(H-,17,18,19,22,23)                                                                                 | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H,19,22)/t15-/m0/s1                                                             | InChI=1S/C17H20BN3O4/c1-11(22)19-9-15-10-21(17(24)25-15)14-3-2-12-4-6-20(16(18)23)7-5-13(12)8-14/h2-3,8,15H,4-7,9-10H2,1H3,(H-,19,22,23)/t15-/m0/s1                                                                  | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,25H,1-2H3,(H,22,23)                                                                                  | InChI=1S/C20H17BN2OS/c1-12-6-18(21)20(22-10-12)23-19-9-15(7-16(11-24)13(19)2)14-4-3-5-17(25)8-14/h3-11,23,25H,1-2H3                                                                                                  | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                          |\n| InChI=1S/C23H24BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3,(H,27,34)                                | InChI=1S/C23H23BFN8O/c1-27-23(34)32-6-4-16(5-7-32)28-12-19-20(24)21(26)33-22(31-19)17(11-30-33)14-8-13-9-15(25)2-3-18(13)29-10-14/h2-3,8-11,16,28H,4-7,12,26H2,1H3/q-1/p+1                                           | Proton(s) added/removed                                                                                                                                                              |\n| InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14H,4,15H2,1-3H3,(H4,20,21,22,23,24)                                                                    | InChI=1S/C19H27N5Si2/c1-4-15-21-19(24-26(3)17-13-9-6-10-14-17)22-18(20)23-25(2)16-11-7-5-8-12-16/h5-14,23H,4,15,20H2,1-3H3,(H,21,22)                                                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H,7,8)(H2,4,5,6)                                                                                                                                     | InChI=1S/CH6BO8P3/c1-11(2,3)9-13(7,8)10-12(4,5)6/h1H3,(H3-,3,4,5,6,7,8)                                                                                                                                              | Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                  |\n| InChI=1S/C7H14BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7H,2-6H2,1H3                                                                                                                                             | InChI=1S/C7H15BNO2S/c1-12(10,11)9-4-2-7(6-8)3-5-9/h7,10H,2-6H2,1H3/q+1                                                                                                                                               | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                      |\n| InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i/hT                                                                                                                                    | InChI=1S/C5H10NO4P/c1-4(7)2-9-11-10-3-5(6)8/h11H,2-3H2,1H3,(H2,6,8)/i11T                                                                                                                                             | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i/hT                                           | InChI=1S/C22H28N4O4/c1-5-29-12-17-25-19-20(26(17)13-22(3,4)28)15-9-7-14(8-10-18(27)30-6-2)11-16(15)24-21(19)23/h7-11,28H,5-6,12-13H2,1-4H3,(H2,23,24)/b10-8+/i28T                                                    | Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                  |\n| InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6,12H,1H3                                                                                                                                | InChI=1S/C9H9BNO4S/c1-7-2-4-8(5-3-7)16(13,14)11-9(12)6-15-10-11/h2-6H,1H3,(H-,12,13)                                                                                                                                 | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)                                                                  |\n| InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H,10,11)                                                                                                                                           | InChI=1S/C7H9NO3/c1-5(9)8-4-2-3-6(8)7(10)11/h1,6H,2-4H2,(H-,9,10,11)                                                                                                                                                 | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C24H25BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18,27H,8-9,11-12,15H2,1H3                     | InChI=1S/C24H24BN6O4S2/c1-36(32,33)19-4-6-20(7-5-19)37(34,35)30-11-8-18(9-12-30)22-13-23(27-15-17-3-2-10-26-14-17)31-24(29-22)21(25)16-28-31/h2-7,10,13-14,16,18H,8-9,11-12,15H2,1H3/q-1/p+1                         | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)                                                      |\n| InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H,34,36)              | InChI=1S/C27H23BF3N4O4S/c1-18-6-8-20(9-7-18)24-16-25(27(29,30)31)33-35(24)21-11-13-22(14-12-21)40(37,38)34-26(36)15-10-19-4-2-3-5-23(19)28-39-17-32/h2-9,11-14,16-17,32H,10,15H2,1H3,(H-,34,36,37)                   | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1/i23+0 | InChI=1S/C22H25FN4O8S/c1-13-11-27(19-9-17(29)18(12-28)35-19)22(32)26(21(13)31)5-7-36(33,34)6-4-25-20(30)14-2-3-16(23)15(8-14)10-24/h2-3,8,11,17-19,28-29H,4-7,9,12H2,1H3,(H,25,30)/t17-,18+,19+/m0/s1                |                                                                                                                                                                                      |\n| InChI=1S/C19H19N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,1-2H3,(H4,20,22,23,24,25,26)                                                      | InChI=1S/C19H18N5O4S/c1-27-13-7-3-11(4-8-13)15-16(12-5-9-14(28-2)10-6-12)22-19-17(21-15)18(20)23-29(25,26)24-19/h3-10,21H,20H2,1-2H3,(H,22,24)/q-1                                                                   | Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)                                          |\n| InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H,22,23)/t15-/m0/s1                                                        | InChI=1S/C20H20BNO5/c1-25-17-8-11-4-7-15(22-10-23)13-9-16(24)14(21)6-5-12(13)18(11)20(27-3)19(17)26-2/h5-6,8-10,15H,4,7H2,1-3H3,(H-,22,23,24)/t15-/m0/s1                                                             | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,21,22,23)/t14-/m1/s1                                                           | InChI=1S/C18H17BN4O/c19-15-8-12-10-21-18(22-13-4-2-1-3-5-13)23-16(12)9-17(15)24-14-6-7-20-11-14/h1-5,8-10,14,20H,6-7,11H2,(H,22,23)/t14-/m1/s1                                                                       | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n| InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11,24H,2,7,10H2,1H3,(H2,23,25,26)                                                           | InChI=1S/C20H19BN2O4/c1-2-7-27-18-6-3-12(8-17(18)24)10-22-11-16-15-9-13(21)4-5-14(15)19(25)23-20(16)26/h3-6,8-9,11H,2,7,10H2,1H3,(H3-,22,23,24,25,26)                                                                | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                           |\n| InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4H,1-3H3,(H,13,17)(H,15,16)                                                                                                            | InChI=1S/C11H12BNO5/c1-5-7(11(2,3)9(15)16)8(14)6(4-18-5)13-10(12)17/h4,17H,1-3H3,(H,15,16)                                                                                                                           | Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)                                                  |\n| InChI=1S/C23H22BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2,(H,30,31)                                                     | InChI=1S/C23H21BN5O2/c24-20-21(14-7-5-13(6-8-14)9-19(30)31)28-23-17(12-27-29(23)22(20)25)16-10-15-3-1-2-4-18(15)26-11-16/h1-4,10-14H,5-9,25H2/q-1/p+1                                                                | Proton(s) added/removed; Omitted undefined stereo                                                                                                                                    |\n| InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,18)                                                                                                  | InChI=1S/C15H18BN3O/c1-15(6-4-2-3-5-7-15)13(20)10-8-17-14-12(10)19-11(16)9-18-14/h8-9H,2-7H2,1H3,(H,17,20)                                                                                                           | Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                           |\n\nDon't know how it is important, but RDKit is complaining about much than those 197 mismatched formulas but produces the same InChI despite of this. The most frequent warning is about undefined stereo - maybe it can be related somehow to poor performance of my network especially identifying stereolayer.\n\n| rdkit warning                                                                                                                                                                                          |   count   &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; |\n|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------:|\n| Omitted undefined stereo                                                                                                                                                                       |              937044 |\n| Accepted unusual valence(s): Si(3); Omitted undefined stereo                                                                                                                                   |                 581 |\n| Accepted unusual valence(s): Si(3)                                                                                                                                                             |                 569 |\n| Accepted unusual valence(s): C(3)                                                                                                                                                              |                 410 |\n| Accepted unusual valence(s): B(2)                                                                                                                                                              |                 300 |\n| Accepted unusual valence(s): C(3); Omitted undefined stereo                                                                                                                                    |                 214 |\n| Accepted unusual valence(s): O(1)                                                                                                                                                              |                 144 |\n| Accepted unusual valence(s): O(1); Omitted undefined stereo                                                                                                                                    |                 121 |\n| Accepted unusual valence(s): B(2); Omitted undefined stereo                                                                                                                                    |                 116 |\n| Accepted unusual valence(s): B(1)                                                                                                                                                              |                 112 |\n| Accepted unusual valence(s): B(1); Omitted undefined stereo                                                                                                                                    |                 112 |\n| Accepted unusual valence(s): N(2)                                                                                                                                                              |                  67 |\n| Accepted unusual valence(s): Si(2)                                                                                                                                                             |                  63 |\n| Charges were rearranged                                                                                                                                                                        |                  54 |\n| Accepted unusual valence(s): N(2); Omitted undefined stereo                                                                                                                                    |                  39 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number)                                                                                                                     |                  39 |\n| Accepted unusual valence(s): Si(2); Omitted undefined stereo                                                                                                                                   |                  37 |\n| Accepted unusual valence(s): C(2); Omitted undefined stereo                                                                                                                                    |                  35 |\n| Accepted unusual valence(s): C(2)                                                                                                                                                              |                  28 |\n| Charges were rearranged; Omitted undefined stereo                                                                                                                                              |                  27 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points)                                                                                    |                  24 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match)                                                                                                |                  20 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2)                                                                                                                                                  |                  14 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number)                                                                                     |                  13 |\n| Accepted unusual valence(s): S(3)                                                                                                                                                              |                  11 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Falsely present, Attachment points)                                                                            |                  10 |\n| Problems/mismatches: Mobile-H( Isotopic: Atoms do not match; Exchangeable isotopic H: Do not match)                                                                                            |                   8 |\n| Accepted unusual valence(s): B+1(1); Omitted undefined stereo                                                                                                                                  |                   6 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Attachment points, Number)                                                                                            |                   5 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: One instead of multiple, Attachment points, Number)                                                            |                   5 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                                        |                   5 |\n| Accepted unusual valence(s): S(3); Omitted undefined stereo                                                                                                                                    |                   4 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match)                                                   |                   4 |\n| Charges were rearranged; Accepted unusual valence(s): O(4); Omitted undefined stereo                                                                                                           |                   4 |\n| Accepted unusual valence(s): B+1(1)                                                                                                                                                            |                   4 |\n| Accepted unusual valence(s): S(1)                                                                                                                                                              |                   4 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2); Omitted undefined stereo                                                                                                                        |                   3 |\n| Charges were rearranged; Accepted unusual valence(s): O(4)                                                                                                                                     |                   3 |\n| Proton(s) added/removed                                                                                                                                                                        |                   3 |\n| Accepted unusual valence(s): C-1(2); C+1(2)                                                                                                                                                    |                   3 |\n| Accepted unusual valence(s): N(1)                                                                                                                                                              |                   3 |\n| Accepted unusual valence(s): C-1(2); C+1(2); Omitted undefined stereo                                                                                                                          |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                                                |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Attachment points, Number; Charge(s): Do not match)                                                    |                   2 |\n| Accepted unusual valence(s): B(1); B+1(1); Omitted undefined stereo                                                                                                                            |                   2 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Charge(s): Do not match; Proton balance: Does not match)                                                                |                   2 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: Attachment points, Number; Proton balance: Does not match)                                                                                     |                   2 |\n| Accepted unusual valence(s): N(1); Omitted undefined stereo                                                                                                                                    |                   1 |\n| Accepted unusual valence(s): I(2)                                                                                                                                                              |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Missing, Attachment points; Proton balance: Does not match)                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number; Mobile-H groups: Attachment points, Number; Proton balance: Does not match)                                                     |                   1 |\n| Accepted unusual valence(s): Si-1(2)                                                                                                                                                           |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si(3)                                                                                                                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match; Proton balance: Does not match)           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Number; Charge(s): Do not match; Proton balance: Does not match)                                   |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number)                                                                                                                                 |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Charge(s): Do not match; Proton balance: Does not match)                                        |                   1 |\n| Accepted unusual valence(s): Si-1(2); Omitted undefined stereo                                                                                                                                 |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Missing, Attachment points; Exchangeable isotopic H: Do not match; Charge(s): Do not match; Proton balance: Does not match) |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Locations or number, Number; Mobile-H groups: Falsely present, Attachment points; Charge(s): Do not match)                                           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: One instead of multiple, Attachment points, Number; Charge(s): Do not match; Proton balance: Does not match)                |                   1 |\n| Accepted unusual valence(s): B(2); B+1(1); Omitted undefined stereo                                                                                                                            |                   1 |\n| Charges were rearranged; Accepted unusual valence(s): C-1(2); C(3); O(4)                                                                                                                       |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si(2); Omitted undefined stereo                                                                                                                          |                   1 |\n| Accepted unusual valence(s): P(2)                                                                                                                                                              |                   1 |\n| Problems/mismatches: Mobile-H( Mobile-H groups: One instead of multiple, Number)                                                                                                               |                   1 |\n| Accepted unusual valence(s): C(3); B+1(1); Omitted undefined stereo                                                                                                                            |                   1 |\n| Proton(s) added/removed; Omitted undefined stereo                                                                                                                                              |                   1 |\n| Accepted unusual valence(s): B(2); B+1(1)                                                                                                                                                      |                   1 |\n| Accepted unusual valence(s): B+1(1); O(1)                                                                                                                                                      |                   1 |\n| Charges were rearranged; Accepted unusual valence(s): C(3); O(4)                                                                                                                               |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si-1(2)                                                                                                                                                  |                   1 |\n| Accepted unusual valence(s): C-1(2); Si+1(2)                                                                                                                                                   |                   1 |\n| Accepted unusual valence(s): C+1(2); Omitted undefined stereo                                                                                                                                  |                   1 |\n| Accepted unusual valence(s): C+1(2); C(2); Omitted undefined stereo                                                                                                                            |                   1 |\n| Accepted unusual valence(s): Si-1(2); Si+1(2); Si(2)                                                                                                                                           |                   1 |\n| Problems/mismatches: Mobile-H( Hydrogens: Number; Mobile-H groups: Multiple instead of one, Attachment points, Number; Charge(s): Do not match)                                                |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si(3)                                                                                                                                                    |                   1 |\n| Accepted unusual valence(s): Si+1(2); Si(2)                                                                                                                                                    |                   1 |\n| Problems/mismatches: Mobile-H( Stereobonds/cumulenes: Extra undefined)                                                                                                                         |                   1 |"
  },
  "source": "meta"
}