change bos/eos/pad to <|endoftext|>

config.json

@@ -1,12 +1,12 @@
 {
-  "_name_or_path": "ahxt/LiteLlama-460M-1T",
+  "_name_or_path": "yuntian-deng/LiteLlama-460M-1T",
   "architectures": [
     "LlamaForCausalLM"
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "bos_token_id": 1,
-  "eos_token_id": 2,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
   "hidden_act": "silu",
   "hidden_size": 1024,
   "initializer_range": 0.02,
@@ -17,7 +17,7 @@
   "num_attention_heads": 16,
   "num_hidden_layers": 24,
   "num_key_value_heads": 2,
-  "pad_token_id": 0,
+  "pad_token_id": 50256,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 1,
-  "eos_token_id": 2,
-  "pad_token_id": 0,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "pad_token_id": 50256,
   "transformers_version": "4.38.0.dev0"
 }

tokenizer.json

@@ -3,24 +3,6 @@
   "truncation": null,
   "padding": null,
   "added_tokens": [
-    {
-      "id": 1,
-      "content": "\"",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": true
-    },
-    {
-      "id": 2,
-      "content": "#",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": true
-    },
     {
       "id": 50256,
       "content": "<|endoftext|>",
@@ -100320,4 +100302,4 @@
       "Ġg azed"
     ]
   }
-}
+}

tokenizer_config.json

@@ -2,22 +2,6 @@
   "add_bos_token": false,
   "add_prefix_space": false,
   "added_tokens_decoder": {
-    "1": {
-      "content": "\"",
-      "lstrip": false,
-      "normalized": true,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "2": {
-      "content": "#",
-      "lstrip": false,
-      "normalized": true,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
     "50256": {
       "content": "<|endoftext|>",
       "lstrip": false,
@@ -29,10 +13,10 @@
   },
   "bos_token": "<|endoftext|>",
   "clean_up_tokenization_spaces": true,
-  "eos_token": "#",
+  "eos_token": "<|endoftext|>",
   "errors": "replace",
   "model_max_length": 1024,
-  "pad_token": "\"",
+  "pad_token": "<|endoftext|>",
   "tokenizer_class": "GPT2Tokenizer",
   "unk_token": "<|endoftext|>"
 }