Update app.py

app.py

@@ -29,17 +29,17 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 peft_model_id = "charansr/llama2-7b-chat-hf-therapist"
 
 config = PeftConfig.from_pretrained(peft_model_id,
-                                   use_auth_token="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", load_in_8bit=True)
+                                   use_auth_token="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", load_in_4bit=True)
 
-newmodel = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_8bit=True,
+newmodel = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_4bit=True,
                                                 use_auth_token="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").to("cpu")
 
 newtokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path,
-                                            use_auth_token="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", load_in_8bit=True).to("cpu")
+                                            use_auth_token="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", load_in_4bit=True).to("cpu")
 
 # Load the Lora model
 newmodel = PeftModel.from_pretrained(newmodel, peft_model_id,
-                                    use_auth_token="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", load_in_8bit=True).to("cpu")
+                                    use_auth_token="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", load_in_4bit=True).to("cpu")
 
 def givetext(input_text,lmodel,ltokenizer):
     try: