fix(tokenizer): Also updates tokenizer.json to prevent any mismatch.
Browse files- tokenizer.json +12 -12
tokenizer.json
CHANGED
@@ -26,9 +26,9 @@
|
|
26 |
"content": "</s>",
|
27 |
"single_word": false,
|
28 |
"lstrip": false,
|
29 |
-
"rstrip":
|
30 |
"normalized": false,
|
31 |
-
"special":
|
32 |
},
|
33 |
{
|
34 |
"id": 32000,
|
@@ -44,7 +44,7 @@
|
|
44 |
"content": "<|assistant|>",
|
45 |
"single_word": false,
|
46 |
"lstrip": false,
|
47 |
-
"rstrip":
|
48 |
"normalized": false,
|
49 |
"special": true
|
50 |
},
|
@@ -53,7 +53,7 @@
|
|
53 |
"content": "<|placeholder1|>",
|
54 |
"single_word": false,
|
55 |
"lstrip": false,
|
56 |
-
"rstrip":
|
57 |
"normalized": false,
|
58 |
"special": true
|
59 |
},
|
@@ -62,7 +62,7 @@
|
|
62 |
"content": "<|placeholder2|>",
|
63 |
"single_word": false,
|
64 |
"lstrip": false,
|
65 |
-
"rstrip":
|
66 |
"normalized": false,
|
67 |
"special": true
|
68 |
},
|
@@ -71,7 +71,7 @@
|
|
71 |
"content": "<|placeholder3|>",
|
72 |
"single_word": false,
|
73 |
"lstrip": false,
|
74 |
-
"rstrip":
|
75 |
"normalized": false,
|
76 |
"special": true
|
77 |
},
|
@@ -80,7 +80,7 @@
|
|
80 |
"content": "<|placeholder4|>",
|
81 |
"single_word": false,
|
82 |
"lstrip": false,
|
83 |
-
"rstrip":
|
84 |
"normalized": false,
|
85 |
"special": true
|
86 |
},
|
@@ -89,7 +89,7 @@
|
|
89 |
"content": "<|system|>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
-
"rstrip":
|
93 |
"normalized": false,
|
94 |
"special": true
|
95 |
},
|
@@ -98,7 +98,7 @@
|
|
98 |
"content": "<|end|>",
|
99 |
"single_word": false,
|
100 |
"lstrip": false,
|
101 |
-
"rstrip":
|
102 |
"normalized": false,
|
103 |
"special": true
|
104 |
},
|
@@ -107,7 +107,7 @@
|
|
107 |
"content": "<|placeholder5|>",
|
108 |
"single_word": false,
|
109 |
"lstrip": false,
|
110 |
-
"rstrip":
|
111 |
"normalized": false,
|
112 |
"special": true
|
113 |
},
|
@@ -116,7 +116,7 @@
|
|
116 |
"content": "<|placeholder6|>",
|
117 |
"single_word": false,
|
118 |
"lstrip": false,
|
119 |
-
"rstrip":
|
120 |
"normalized": false,
|
121 |
"special": true
|
122 |
},
|
@@ -125,7 +125,7 @@
|
|
125 |
"content": "<|user|>",
|
126 |
"single_word": false,
|
127 |
"lstrip": false,
|
128 |
-
"rstrip":
|
129 |
"normalized": false,
|
130 |
"special": true
|
131 |
}
|
|
|
26 |
"content": "</s>",
|
27 |
"single_word": false,
|
28 |
"lstrip": false,
|
29 |
+
"rstrip": true,
|
30 |
"normalized": false,
|
31 |
+
"special": false
|
32 |
},
|
33 |
{
|
34 |
"id": 32000,
|
|
|
44 |
"content": "<|assistant|>",
|
45 |
"single_word": false,
|
46 |
"lstrip": false,
|
47 |
+
"rstrip": true,
|
48 |
"normalized": false,
|
49 |
"special": true
|
50 |
},
|
|
|
53 |
"content": "<|placeholder1|>",
|
54 |
"single_word": false,
|
55 |
"lstrip": false,
|
56 |
+
"rstrip": true,
|
57 |
"normalized": false,
|
58 |
"special": true
|
59 |
},
|
|
|
62 |
"content": "<|placeholder2|>",
|
63 |
"single_word": false,
|
64 |
"lstrip": false,
|
65 |
+
"rstrip": true,
|
66 |
"normalized": false,
|
67 |
"special": true
|
68 |
},
|
|
|
71 |
"content": "<|placeholder3|>",
|
72 |
"single_word": false,
|
73 |
"lstrip": false,
|
74 |
+
"rstrip": true,
|
75 |
"normalized": false,
|
76 |
"special": true
|
77 |
},
|
|
|
80 |
"content": "<|placeholder4|>",
|
81 |
"single_word": false,
|
82 |
"lstrip": false,
|
83 |
+
"rstrip": true,
|
84 |
"normalized": false,
|
85 |
"special": true
|
86 |
},
|
|
|
89 |
"content": "<|system|>",
|
90 |
"single_word": false,
|
91 |
"lstrip": false,
|
92 |
+
"rstrip": true,
|
93 |
"normalized": false,
|
94 |
"special": true
|
95 |
},
|
|
|
98 |
"content": "<|end|>",
|
99 |
"single_word": false,
|
100 |
"lstrip": false,
|
101 |
+
"rstrip": true,
|
102 |
"normalized": false,
|
103 |
"special": true
|
104 |
},
|
|
|
107 |
"content": "<|placeholder5|>",
|
108 |
"single_word": false,
|
109 |
"lstrip": false,
|
110 |
+
"rstrip": true,
|
111 |
"normalized": false,
|
112 |
"special": true
|
113 |
},
|
|
|
116 |
"content": "<|placeholder6|>",
|
117 |
"single_word": false,
|
118 |
"lstrip": false,
|
119 |
+
"rstrip": true,
|
120 |
"normalized": false,
|
121 |
"special": true
|
122 |
},
|
|
|
125 |
"content": "<|user|>",
|
126 |
"single_word": false,
|
127 |
"lstrip": false,
|
128 |
+
"rstrip": true,
|
129 |
"normalized": false,
|
130 |
"special": true
|
131 |
}
|