shaipranesh commited on
Commit
ae118b1
·
verified ·
1 Parent(s): 4f25fbc

Upload five finalized 20M ligand shards

Browse files
.gitattributes CHANGED
@@ -58,3 +58,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
+ shard_1/train.lmdb filter=lfs diff=lfs merge=lfs -text
62
+ shard_2/train.lmdb filter=lfs diff=lfs merge=lfs -text
63
+ shard_3/train.lmdb filter=lfs diff=lfs merge=lfs -text
64
+ shard_4/train.lmdb filter=lfs diff=lfs merge=lfs -text
65
+ shard_5/train.lmdb filter=lfs diff=lfs merge=lfs -text
shard_1/metadata.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_registry_unique": 19999982,
3
+ "base_seen_lmdb": "/scratch/shai/data_20m_selfies_original_cartesian_shards/selfies_seen.lmdb",
4
+ "conformers_per_molecule": 10,
5
+ "coordinate_contract": "coordinates[:, i, :] matches atom i after SELFIES decode and RDKit parse",
6
+ "coordinate_space": "original RDKit Cartesian coordinates; atom-axis permutation only",
7
+ "discarded_after_split_filled": 0,
8
+ "duplicate_selfies_in_base_rejected": 6,
9
+ "duplicate_selfies_in_shard_rejected": 9158,
10
+ "elapsed_seconds": 8426.422278642654,
11
+ "finished_at_epoch": 1787190989.081326,
12
+ "local_registry_digest_entries": 20000000,
13
+ "local_registry_extra_entries_from_resume": 0,
14
+ "local_registry_unique": 20000000,
15
+ "molecular_property_contract": {
16
+ "calculated_from": "canonical heavy-atom molecule represented by smi; standard implicit hydrogens are included by RDKit",
17
+ "molecular_weight_kind": "average molecular weight, not exact isotope mass",
18
+ "rdkit_version": "2025.03.6",
19
+ "sa_score_implementation": "RDKit Contrib SA_Score/sascorer.py"
20
+ },
21
+ "partition_count": 5,
22
+ "partition_index": 0,
23
+ "partition_rejected": 6796319,
24
+ "raw_slices": [
25
+ 1,
26
+ 2,
27
+ 3,
28
+ 4,
29
+ 5
30
+ ],
31
+ "record_format": {
32
+ "atoms": "heavy-atom symbols in decoded SELFIES atom order",
33
+ "coordinates": "float32 ndarray, shape (10, n_heavy, 3)",
34
+ "mol_log_p": "float; RDKit Crippen.MolLogP",
35
+ "molecular_weight": "float; RDKit Descriptors.MolWt",
36
+ "qed": "float; RDKit QED.qed",
37
+ "sa_score": "float; RDKit Contrib SA_Score; 1 is easiest and 10 is hardest",
38
+ "smi": "SELFIES encoding of canonical isomeric heavy-atom SMILES",
39
+ "tpsa": "float; RDKit rdMolDescriptors.CalcTPSA"
40
+ },
41
+ "resume": {
42
+ "enabled": true,
43
+ "resumed_from_train": 18309263,
44
+ "resumed_from_val": 0,
45
+ "source_records_skipped": 91564160
46
+ },
47
+ "skipped_by_reason": {
48
+ "coordinates_shape": 6638
49
+ },
50
+ "source_file_error_samples": [],
51
+ "source_file_errors": {},
52
+ "source_processed": {
53
+ "raw_zinc": 8502912
54
+ },
55
+ "source_written": {
56
+ "raw_zinc": 1690737
57
+ },
58
+ "split": "all accepted records assigned to train",
59
+ "started_at_epoch": 1787182562.6590474,
60
+ "total_examples": 20000000,
61
+ "train_examples": 20000000,
62
+ "val_examples": 0,
63
+ "validation": {
64
+ "selfies_roundtrip_every": 10000,
65
+ "smoke_test": false
66
+ },
67
+ "workers": 192
68
+ }
shard_1/train.lmdb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:499368580f8dacc2fb844b867c8c1695f5d5a4aaffa2314122fea04191b49e9a
3
+ size 86753845248
shard_2/metadata.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_registry_unique": 19999982,
3
+ "base_seen_lmdb": "/scratch/shai/data_20m_selfies_original_cartesian_shards/selfies_seen.lmdb",
4
+ "conformers_per_molecule": 10,
5
+ "coordinate_contract": "coordinates[:, i, :] matches atom i after SELFIES decode and RDKit parse",
6
+ "coordinate_space": "original RDKit Cartesian coordinates; atom-axis permutation only",
7
+ "discarded_after_split_filled": 0,
8
+ "duplicate_selfies_in_base_rejected": 23,
9
+ "duplicate_selfies_in_shard_rejected": 68,
10
+ "elapsed_seconds": 84522.61968636513,
11
+ "finished_at_epoch": 1787110423.9621181,
12
+ "local_registry_unique": 20000000,
13
+ "molecular_property_contract": {
14
+ "calculated_from": "canonical heavy-atom molecule represented by smi; standard implicit hydrogens are included by RDKit",
15
+ "molecular_weight_kind": "average molecular weight, not exact isotope mass",
16
+ "rdkit_version": "2025.03.6",
17
+ "sa_score_implementation": "RDKit Contrib SA_Score/sascorer.py"
18
+ },
19
+ "partition_count": 5,
20
+ "partition_index": 1,
21
+ "partition_rejected": 79990241,
22
+ "raw_slices": [
23
+ 2,
24
+ 3,
25
+ 4,
26
+ 5,
27
+ 1
28
+ ],
29
+ "record_format": {
30
+ "atoms": "heavy-atom symbols in decoded SELFIES atom order",
31
+ "coordinates": "float32 ndarray, shape (10, n_heavy, 3)",
32
+ "mol_log_p": "float; RDKit Crippen.MolLogP",
33
+ "molecular_weight": "float; RDKit Descriptors.MolWt",
34
+ "qed": "float; RDKit QED.qed",
35
+ "sa_score": "float; RDKit Contrib SA_Score; 1 is easiest and 10 is hardest",
36
+ "smi": "SELFIES encoding of canonical isomeric heavy-atom SMILES",
37
+ "tpsa": "float; RDKit rdMolDescriptors.CalcTPSA"
38
+ },
39
+ "skipped_by_reason": {
40
+ "coordinates_shape": 77717,
41
+ "disconnected_heavy_graph": 1
42
+ },
43
+ "source_file_error_samples": [],
44
+ "source_file_errors": {},
45
+ "source_processed": {
46
+ "raw_zinc": 100068096
47
+ },
48
+ "source_written": {
49
+ "raw_zinc": 20000000
50
+ },
51
+ "split": "all accepted records assigned to train",
52
+ "started_at_epoch": 1787025901.3424318,
53
+ "total_examples": 20000000,
54
+ "train_examples": 20000000,
55
+ "val_examples": 0,
56
+ "validation": {
57
+ "selfies_roundtrip_every": 10000,
58
+ "smoke_test": false
59
+ },
60
+ "workers": 192
61
+ }
shard_2/train.lmdb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a148d129f64ad5c50fae4edf6def3439e6c2083da3c297428c68114b6d7b112b
3
+ size 86744469504
shard_3/metadata.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_registry_unique": 19999982,
3
+ "base_seen_lmdb": "/scratch/shai/data_20m_selfies_original_cartesian_shards/selfies_seen.lmdb",
4
+ "conformers_per_molecule": 10,
5
+ "coordinate_contract": "coordinates[:, i, :] matches atom i after SELFIES decode and RDKit parse",
6
+ "coordinate_space": "original RDKit Cartesian coordinates; atom-axis permutation only",
7
+ "discarded_after_split_filled": 0,
8
+ "duplicate_selfies_in_base_rejected": 2,
9
+ "duplicate_selfies_in_shard_rejected": 59626,
10
+ "elapsed_seconds": 5763.920112133026,
11
+ "finished_at_epoch": 1787188326.67951,
12
+ "local_registry_digest_entries": 20000000,
13
+ "local_registry_extra_entries_from_resume": 0,
14
+ "local_registry_unique": 20000000,
15
+ "molecular_property_contract": {
16
+ "calculated_from": "canonical heavy-atom molecule represented by smi; standard implicit hydrogens are included by RDKit",
17
+ "molecular_weight_kind": "average molecular weight, not exact isotope mass",
18
+ "rdkit_version": "2025.03.6",
19
+ "sa_score_implementation": "RDKit Contrib SA_Score/sascorer.py"
20
+ },
21
+ "partition_count": 5,
22
+ "partition_index": 2,
23
+ "partition_rejected": 4404048,
24
+ "raw_slices": [
25
+ 3,
26
+ 4,
27
+ 5,
28
+ 1,
29
+ 2
30
+ ],
31
+ "record_format": {
32
+ "atoms": "heavy-atom symbols in decoded SELFIES atom order",
33
+ "coordinates": "float32 ndarray, shape (10, n_heavy, 3)",
34
+ "mol_log_p": "float; RDKit Crippen.MolLogP",
35
+ "molecular_weight": "float; RDKit Descriptors.MolWt",
36
+ "qed": "float; RDKit QED.qed",
37
+ "sa_score": "float; RDKit Contrib SA_Score; 1 is easiest and 10 is hardest",
38
+ "smi": "SELFIES encoding of canonical isomeric heavy-atom SMILES",
39
+ "tpsa": "float; RDKit rdMolDescriptors.CalcTPSA"
40
+ },
41
+ "resume": {
42
+ "enabled": true,
43
+ "resumed_from_train": 18959723,
44
+ "resumed_from_val": 0,
45
+ "source_records_skipped": 94505600
46
+ },
47
+ "skipped_by_reason": {},
48
+ "source_file_error_samples": [],
49
+ "source_file_errors": {},
50
+ "source_processed": {
51
+ "raw_zinc": 5504000
52
+ },
53
+ "source_written": {
54
+ "raw_zinc": 1040277
55
+ },
56
+ "split": "all accepted records assigned to train",
57
+ "started_at_epoch": 1787182562.759398,
58
+ "total_examples": 20000000,
59
+ "train_examples": 20000000,
60
+ "val_examples": 0,
61
+ "validation": {
62
+ "selfies_roundtrip_every": 10000,
63
+ "smoke_test": false
64
+ },
65
+ "workers": 192
66
+ }
shard_3/train.lmdb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2816d4c72a03923787240ff8672bf465de43af3198d8aacd6605664e36a32144
3
+ size 86745681920
shard_4/metadata.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_registry_unique": 19999982,
3
+ "base_seen_lmdb": "/scratch/shai/data_20m_selfies_original_cartesian_shards/selfies_seen.lmdb",
4
+ "conformers_per_molecule": 10,
5
+ "coordinate_contract": "coordinates[:, i, :] matches atom i after SELFIES decode and RDKit parse",
6
+ "coordinate_space": "original RDKit Cartesian coordinates; atom-axis permutation only",
7
+ "discarded_after_split_filled": 0,
8
+ "duplicate_selfies_in_base_rejected": 6,
9
+ "duplicate_selfies_in_shard_rejected": 83218,
10
+ "elapsed_seconds": 20922.117764234543,
11
+ "finished_at_epoch": 1787203484.878445,
12
+ "local_registry_digest_entries": 20000000,
13
+ "local_registry_extra_entries_from_resume": 0,
14
+ "local_registry_unique": 20000000,
15
+ "molecular_property_contract": {
16
+ "calculated_from": "canonical heavy-atom molecule represented by smi; standard implicit hydrogens are included by RDKit",
17
+ "molecular_weight_kind": "average molecular weight, not exact isotope mass",
18
+ "rdkit_version": "2025.03.6",
19
+ "sa_score_implementation": "RDKit Contrib SA_Score/sascorer.py"
20
+ },
21
+ "partition_count": 5,
22
+ "partition_index": 3,
23
+ "partition_rejected": 14806742,
24
+ "raw_slices": [
25
+ 4,
26
+ 5,
27
+ 1,
28
+ 2,
29
+ 3
30
+ ],
31
+ "record_format": {
32
+ "atoms": "heavy-atom symbols in decoded SELFIES atom order",
33
+ "coordinates": "float32 ndarray, shape (10, n_heavy, 3)",
34
+ "mol_log_p": "float; RDKit Crippen.MolLogP",
35
+ "molecular_weight": "float; RDKit Descriptors.MolWt",
36
+ "qed": "float; RDKit QED.qed",
37
+ "sa_score": "float; RDKit Contrib SA_Score; 1 is easiest and 10 is hardest",
38
+ "smi": "SELFIES encoding of canonical isomeric heavy-atom SMILES",
39
+ "tpsa": "float; RDKit rdMolDescriptors.CalcTPSA"
40
+ },
41
+ "resume": {
42
+ "enabled": true,
43
+ "resumed_from_train": 16383299,
44
+ "resumed_from_val": 0,
45
+ "source_records_skipped": 81510912
46
+ },
47
+ "skipped_by_reason": {
48
+ "disconnected_heavy_graph": 1
49
+ },
50
+ "source_file_error_samples": [
51
+ {
52
+ "error": "Error: /scratch/shai/data_new_hydrogens/4/1/28/63600.lmdb: Bad file descriptor",
53
+ "path": "/scratch/shai/data_new_hydrogens/4/1/28/63600.lmdb"
54
+ },
55
+ {
56
+ "error": "Error: /scratch/shai/data_new_hydrogens/4/3/57/63600.lmdb: Bad file descriptor",
57
+ "path": "/scratch/shai/data_new_hydrogens/4/3/57/63600.lmdb"
58
+ }
59
+ ],
60
+ "source_file_errors": {
61
+ "Error": 2
62
+ },
63
+ "source_processed": {
64
+ "raw_zinc": 18506752
65
+ },
66
+ "source_written": {
67
+ "raw_zinc": 3616701
68
+ },
69
+ "split": "all accepted records assigned to train",
70
+ "started_at_epoch": 1787182562.7606807,
71
+ "total_examples": 20000000,
72
+ "train_examples": 20000000,
73
+ "val_examples": 0,
74
+ "validation": {
75
+ "selfies_roundtrip_every": 10000,
76
+ "smoke_test": false
77
+ },
78
+ "workers": 192
79
+ }
shard_4/train.lmdb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af30888477592ab8fdca9c3684308b6a4ce60bcac010db0e7cb298045eeffd70
3
+ size 86746476544
shard_5/metadata.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_registry_unique": 19999982,
3
+ "base_seen_lmdb": "/scratch/shai/data_20m_selfies_original_cartesian_shards/selfies_seen.lmdb",
4
+ "conformers_per_molecule": 10,
5
+ "coordinate_contract": "coordinates[:, i, :] matches atom i after SELFIES decode and RDKit parse",
6
+ "coordinate_space": "original RDKit Cartesian coordinates; atom-axis permutation only",
7
+ "discarded_after_split_filled": 0,
8
+ "duplicate_selfies_in_base_rejected": 5,
9
+ "duplicate_selfies_in_shard_rejected": 8350,
10
+ "elapsed_seconds": 17794.491253614426,
11
+ "finished_at_epoch": 1787200396.49274,
12
+ "local_registry_digest_entries": 20000000,
13
+ "local_registry_extra_entries_from_resume": 0,
14
+ "local_registry_unique": 20000000,
15
+ "molecular_property_contract": {
16
+ "calculated_from": "canonical heavy-atom molecule represented by smi; standard implicit hydrogens are included by RDKit",
17
+ "molecular_weight_kind": "average molecular weight, not exact isotope mass",
18
+ "rdkit_version": "2025.03.6",
19
+ "sa_score_implementation": "RDKit Contrib SA_Score/sascorer.py"
20
+ },
21
+ "partition_count": 5,
22
+ "partition_index": 4,
23
+ "partition_rejected": 13196282,
24
+ "raw_slices": [
25
+ 5,
26
+ 1,
27
+ 2,
28
+ 3,
29
+ 4
30
+ ],
31
+ "record_format": {
32
+ "atoms": "heavy-atom symbols in decoded SELFIES atom order",
33
+ "coordinates": "float32 ndarray, shape (10, n_heavy, 3)",
34
+ "mol_log_p": "float; RDKit Crippen.MolLogP",
35
+ "molecular_weight": "float; RDKit Descriptors.MolWt",
36
+ "qed": "float; RDKit QED.qed",
37
+ "sa_score": "float; RDKit Contrib SA_Score; 1 is easiest and 10 is hardest",
38
+ "smi": "SELFIES encoding of canonical isomeric heavy-atom SMILES",
39
+ "tpsa": "float; RDKit rdMolDescriptors.CalcTPSA"
40
+ },
41
+ "resume": {
42
+ "enabled": true,
43
+ "resumed_from_train": 16708446,
44
+ "resumed_from_val": 0,
45
+ "source_records_skipped": 83496064
46
+ },
47
+ "skipped_by_reason": {
48
+ "disconnected_heavy_graph": 1
49
+ },
50
+ "source_file_error_samples": [
51
+ {
52
+ "error": "Error: /scratch/shai/data_new_hydrogens/5/2/96/61100.lmdb: Bad file descriptor",
53
+ "path": "/scratch/shai/data_new_hydrogens/5/2/96/61100.lmdb"
54
+ },
55
+ {
56
+ "error": "Error: /scratch/shai/data_new_hydrogens/5/5/157/61500.lmdb: Bad file descriptor",
57
+ "path": "/scratch/shai/data_new_hydrogens/5/5/157/61500.lmdb"
58
+ }
59
+ ],
60
+ "source_file_errors": {
61
+ "Error": 2
62
+ },
63
+ "source_processed": {
64
+ "raw_zinc": 16496256
65
+ },
66
+ "source_written": {
67
+ "raw_zinc": 3291554
68
+ },
69
+ "split": "all accepted records assigned to train",
70
+ "started_at_epoch": 1787182602.0014863,
71
+ "total_examples": 20000000,
72
+ "train_examples": 20000000,
73
+ "val_examples": 0,
74
+ "validation": {
75
+ "selfies_roundtrip_every": 10000,
76
+ "smoke_test": false
77
+ },
78
+ "workers": 192
79
+ }
shard_5/train.lmdb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3eda8b36efad3464bcd0239865857806c1d8ccfc8dab72e8c1abccbc71482b5
3
+ size 86750724096