Upload tokenizer

added_tokens.json

@@ -1,4 +1,5 @@
 {
-  "<|im_end|>": 32000,
-  "<|im_start|>": 32001
+  "<sep>": 32002,
+  "<|end_of_turn|>": 32000,
+  "<|pad_0|>": 32001
 }

special_tokens_map.json

@@ -1,8 +1,7 @@
 {
   "additional_special_tokens": [
-    "<unk>",
-    "<s>",
-    "</s>"
+    "<|end_of_turn|>",
+    "<|pad_0|>"
   ],
   "bos_token": {
     "content": "<s>",
@@ -12,7 +11,21 @@
     "single_word": false
   },
   "eos_token": {
-    "content": "</s>",
+    "content": "<|end_of_turn|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|end_of_turn|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "<sep>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer.json

@@ -32,7 +32,7 @@
     },
     {
       "id": 32000,
-      "content": "<|im_end|>",
+      "content": "<|end_of_turn|>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -41,7 +41,16 @@
     },
     {
       "id": 32001,
-      "content": "<|im_start|>",
+      "content": "<|pad_0|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 32002,
+      "content": "<sep>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,

tokenizer_config.json

@@ -25,7 +25,7 @@
       "special": true
     },
     "32000": {
-      "content": "<|im_end|>",
+      "content": "<|end_of_turn|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -33,7 +33,15 @@
       "special": true
     },
     "32001": {
-      "content": "<|im_start|>",
+      "content": "<|pad_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32002": {
+      "content": "<sep>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -42,19 +50,18 @@
     }
   },
   "additional_special_tokens": [
-    "<unk>",
-    "<s>",
-    "</s>"
+    "<|end_of_turn|>",
+    "<|pad_0|>"
   ],
   "bos_token": "<s>",
+  "chat_template": "{{ bos_token }}{% for message in messages %}{{ 'GPT4 Correct ' + message['role'].title() + ': ' + message['content'] + '<|end_of_turn|>'}}{% endfor %}{% if add_generation_prompt %}{{ 'GPT4 Correct Assistant:' }}{% endif %}",
   "clean_up_tokenization_spaces": false,
-  "device_map": {
-    "": "cuda"
-  },
-  "eos_token": "</s>",
+  "device_map": "auto",
+  "eos_token": "<|end_of_turn|>",
   "legacy": true,
   "model_max_length": 1000000000000000019884624838656,
-  "pad_token": null,
+  "pad_token": "<|end_of_turn|>",
+  "sep_token": "<sep>",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,
   "tokenizer_class": "LlamaTokenizer",