Update app.py

app.py

@@ -5,7 +5,6 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 import gradio as gr
 import os
 READ_HF=os.environ["read_hf"]
-
 from unsloth import FastLanguageModel
 
 
@@ -74,7 +73,7 @@ def chunk_it(inventory_list, user_input_text):
     model, tokenizer = FastLanguageModel.from_pretrained(
     model_name = "VanguardAI/CoT_multi_llama_LoRA_4bit", # YOUR MODEL YOU USED FOR TRAINING
     max_seq_length = 2048,
-    dtype = None,
+    dtype = torch.bfloat16,
     load_in_4bit = True,
     token= READ_HF    
     )