Fix: Environment issues resolved, full Fast Tokenizer uploaded.

chat_template.jinja

@@ -0,0 +1 @@
+{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}

tokenizer_config.json

@@ -4,10 +4,11 @@
   "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
   "eos_token": "</s>",
-  "model_file": "tmp_factory/tokenizer_en/spm.model",
+  "is_local": true,
   "model_max_length": 1000000000000000019884624838656,
   "pad_token": "<pad>",
-  "tokenizer_class": "LlamaTokenizer",
+  "tokenizer_class": "TokenizersBackend",
   "unk_token": "<unk>",
-  "use_default_system_prompt": false
+  "use_default_system_prompt": false,
+  "use_fast": true
 }