switched to shared tokenizer

app.py

@@ -9,8 +9,9 @@ instruct_model_path = "MaxBlumenfeld/smollm2-135m-bootleg-instruct"
 
 
 base_tokenizer = AutoTokenizer.from_pretrained(base_model_id)
-# instruct_tokenizer = AutoTokenizer.from_pretrained(instruct_model_path, local_files_only=True)
-instruct_tokenizer = AutoTokenizer.from_pretrained(instruct_model_path)
+# instruct_tokenizer = AutoTokenizer.from_pretrained(instruct_model_path)
+
+
 
 base_model = AutoModelForCausalLM.from_pretrained(base_model_id)
 # instruct_model = AutoModelForCausalLM.from_pretrained(instruct_model_path, local_files_only=True)
@@ -68,7 +69,7 @@ def chat(message, temperature, max_length, system_prompt):
     
     instruct_response = generate_response(
         instruct_model, 
-        instruct_tokenizer, 
+        base_tokenizer, 
         message, 
         temperature, 
         max_length,