Fixed tokenizer (hopefully)

[SYSTEM_PROMPT] and [/SYSTEM_PROMPT] moved from token IDs 131072/131073 to 17/18, as was originally intended. Extraneous tokens removed from config. Embedding weights and lm_head altered with the Python Transformers toolkit to accommodate these changes.

config.json

@@ -23,5 +23,5 @@
   "torch_dtype": "bfloat16",
   "transformers_version": "4.49.0.dev0",
   "use_cache": false,
-  "vocab_size": 131074
+  "vocab_size": 131072
 }

model-00001-of-00005.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a131f81d2b31fefbe344c904e613ef8161baee5036a59888d2b0ab523ad165ec
-size 4865542976
+oid sha256:741eb08fd622a4cc8331a59fc7262f9e61f1a53f24d439e6a820bb2ea3c22422
+size 4865522464

model-00005-of-00005.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4794b542a546b181bcc7708a04c4c70d278e13cec5abd20cd74fe12e9b228369
-size 4907516752
+oid sha256:34e82af09f94e8a3f8c9959f7623fe404660e2e6eb56a72ddffbf1d859e031dd
+size 4907496240

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a26d19ba141138e260796b82a35d1e4a783d007706820702ba03ec160c3b2efc
-size 17078679
+oid sha256:38b3cc4f6ed44a0c1e023d21fcb148b555d288cb74dac642eb49bdfe58a3a70e
+size 21525621

tokenizer_config.json

@@ -140,7 +140,7 @@
       "special": true
     },
     "17": {
-      "content": "<SPECIAL_17>",
+      "content": "[SYSTEM_PROMPT]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -148,7 +148,7 @@
       "special": true
     },
     "18": {
-      "content": "<SPECIAL_18>",
+      "content": "[/SYSTEM_PROMPT]",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -8002,22 +8002,6 @@
       "rstrip": false,
       "single_word": false,
       "special": true
-    },
-    "131072": {
-      "content": "[SYSTEM_PROMPT]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "131073": {
-      "content": "[/SYSTEM_PROMPT]",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
     }
   },
   "bos_token": "<s>",
@@ -8031,6 +8015,6 @@
   ],
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<pad>",
-  "tokenizer_class": "PreTrainedTokenizer",
+  "tokenizer_class": "PreTrainedTokenizerFast",
   "unk_token": "<unk>"
 }