modded to work with llama.cpp_HF and 72b

it works, hooray. original is in previous commit

config.json

@@ -5,7 +5,7 @@
   "attention_dropout": 0.0,
   "bos_token_id": 151643,
   "decoder_sparse_step": 1,
-  "eos_token_id": 151643,
+  "eos_token_id": 151645,
   "hidden_act": "silu",
   "hidden_size": 3584,
   "initializer_range": 0.02,

special_tokens_map.json

@@ -1,6 +1,6 @@
 {
   "eos_token": {
-    "content": "<|endoftext|>",
+    "content": "<|im_end|>",
     "lstrip": false,
     "normalized": false,
     "rstrip": false,

tokenizer_config.json

@@ -30,7 +30,7 @@
   "bos_token": null,
   "chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|endoftext|>",
+  "eos_token": "<|im_end|>",
   "errors": "replace",
   "model_max_length": 32768,
   "pad_token": "<|endoftext|>",