Update app.py

app.py

@@ -7,16 +7,19 @@ import torch
 import transformers
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
+# App framework
+st.title('🦜Seon\'s Legal QA For Dummies 🔗 ')
+prompt = PromptTemplate(template=template, input_variables=["text"])
 
-offload_folder = 'C:\model_weights'
 
+offload_folder = 'C:\model_weights'
 
 model = AutoModelForCausalLM.from_pretrained("PyaeSoneK/LlamaV2LegalFineTuned",
                                     device_map='auto',
                                   torch_dtype=torch.float16,
                                   use_auth_token= st.secrets['hf_access_token'],
                                     offload_folder=offload_folder,
-                                    low_cpu_mem_usage=True)
+                                    )
                                             #  load_in_4bit=True
 
 tokenizer = AutoTokenizer.from_pretrained("PyaeSoneK/LlamaV2LegalFineTuned",
@@ -103,14 +106,16 @@ template = get_prompt(instruction, system_prompt)
 print(template)
 
 
-# App framework
-st.title('🦜Seon\'s Legal QA For Dummies 🔗 ')
-prompt = PromptTemplate(template=template, input_variables=["text"])
+
 llm_chain = LLMChain(prompt=prompt, llm=llm)
 
 
 text = st.text_input('Plug in your prompt here') 
 # Instantiate the prompt template # this will show stuff to the screen if there's a prompt 
+
 if text:
     response = llm_chain.run(text)
-    st.write(parse_text(response)) 
+    st.write(parse_text(response)) 
+
+
+