Update app.py

app.py

@@ -3,10 +3,9 @@ import gradio as gr
 
 from peft import PeftModel, PeftConfig
 from transformers import AutoModelForCausalLM, AutoTokenizer ,pipeline, BitsAndBytesConfig
-
-#config = PeftConfig.from_pretrained("ShishuTripathi/entity_coder")
-#model = AutoModelForCausalLM.from_pretrained("ybelkada/falcon-7b-sharded-bf16",trust_remote_code=True)
-model = AutoModelForCausalLM.from_pretrained("ShishuTripathi/entity_coder")
+config = PeftConfig.from_pretrained("ShishuTripathi/entity_coder")
+model = AutoModelForCausalLM.from_pretrained("ybelkada/falcon-7b-sharded-bf16",trust_remote_code=True,torch_dtype=torch.bfloat16,load_in_8bit=True,device_map={"":"cpu"},)
+model = PeftModel.from_pretrained(model,"ShishuTripathi/entity_coder")
 tokenizer = AutoTokenizer.from_pretrained("ShishuTripathi/entity_coder")
 generator = pipeline('text-generation' , model = model, tokenizer =tokenizer, max_length = 50)