SagiPolaczek
commited on
Commit
•
f8371c4
1
Parent(s):
9b8f2d0
Push model using huggingface_hub.
Browse files
tokenizer/bpe_tokenizer_trained_on_chembl_zinc_with_aug_4272372_samples_balanced_1_1.json
CHANGED
@@ -86,7 +86,7 @@
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
-
"content": "<
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
@@ -2853,7 +2853,7 @@
|
|
2853 |
"<MOLECULAR_ENTITY>": 6,
|
2854 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2855 |
"<INTERNAL_0>": 8,
|
2856 |
-
"<
|
2857 |
"<INTERNAL_2>": 10,
|
2858 |
"<INTERNAL_3>": 11,
|
2859 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
+
"content": "<MOLECULAR_ENTITY_EPITOPE>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
|
|
2853 |
"<MOLECULAR_ENTITY>": 6,
|
2854 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2855 |
"<INTERNAL_0>": 8,
|
2856 |
+
"<MOLECULAR_ENTITY_EPITOPE>": 9,
|
2857 |
"<INTERNAL_2>": 10,
|
2858 |
"<INTERNAL_3>": 11,
|
2859 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
tokenizer/cell_attributes_tokenizer.json
CHANGED
@@ -86,7 +86,7 @@
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
-
"content": "<
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
@@ -2858,7 +2858,7 @@
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
-
"<
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
+
"content": "<MOLECULAR_ENTITY_EPITOPE>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
+
"<MOLECULAR_ENTITY_EPITOPE>": 9,
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
tokenizer/gene_tokenizer.json
CHANGED
@@ -86,7 +86,7 @@
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
-
"content": "<
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
@@ -2858,7 +2858,7 @@
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
-
"<
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
+
"content": "<MOLECULAR_ENTITY_EPITOPE>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
+
"<MOLECULAR_ENTITY_EPITOPE>": 9,
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
tokenizer/t5_tokenizer_AA_special.json
CHANGED
@@ -86,7 +86,7 @@
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
-
"content": "<
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
@@ -2858,7 +2858,7 @@
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
-
"<
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|
|
|
86 |
},
|
87 |
{
|
88 |
"id": 9,
|
89 |
+
"content": "<MOLECULAR_ENTITY_EPITOPE>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
"rstrip": false,
|
|
|
2858 |
"<MOLECULAR_ENTITY>": 6,
|
2859 |
"<GLOBAL_INTERACTION_ATTRIBUTES>": 7,
|
2860 |
"<INTERNAL_0>": 8,
|
2861 |
+
"<MOLECULAR_ENTITY_EPITOPE>": 9,
|
2862 |
"<INTERNAL_2>": 10,
|
2863 |
"<INTERNAL_3>": 11,
|
2864 |
"<MOLECULAR_ENTITY_TCR_ALPHA_CHAIN>": 12,
|