Update app.py

app.py

@@ -25,13 +25,15 @@ 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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")
+                                   use_auth_token="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", load_in_8bit=True, device_map='auto',)
+
 newmodel = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_8bit=True, device_map='auto',
                                                 use_auth_token="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")
 
 newtokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path,
-                                            use_auth_token="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")
+                                            use_auth_token="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", load_in_8bit=True, device_map='auto',)
 
 # Load the Lora model
 newmodel = PeftModel.from_pretrained(newmodel, peft_model_id,