Update fn.py

fn.py

@@ -203,29 +203,6 @@ def chat(message, history = [], instruction = None, args = {}):
 
     return model_output
 
-def infer(args: dict):
-    global cfg
-
-    if 'model_name' in args:
-        load_model(args['model_name'], args['qtype'], args['dtype'])
-
-    for k in [
-        'instruction',
-        'inst_template',
-        'chat_template',
-        'max_new_tokens',
-        'temperature',
-        'top_p',
-        'top_k',
-        'repetition_penalty'
-        ]:
-        cfg[k] = args[k]
-
-    if 'messages' in args:
-        return chat(args['input'], args['messages'])
-    if 'instruction' in args:
-        return instruct(args['instruction'], args['input'])
-
 def apply_template(messages):
     global tokenizer, cfg
 
@@ -235,6 +212,6 @@ def apply_template(messages):
     if type(messages) is str:
         if cfg['inst_template']:
             return cfg['inst_template'].format(instruction=cfg['instruction'], input=messages)
-        return cfg['instruction']
+        return cfg['instruction'].format(input=messages)
     if type(messages) is list:
         return tokenizer.apply_chat_template(conversation=messages, add_generation_prompt=True, tokenize=False)