Update app.py

app.py

@@ -32,6 +32,8 @@ model = model = GPT4All(model_name, model_path, allow_download = False, device="
 
 
 # creating a pdf reader object
+
+"""
 reader = PdfReader("./resource/NGAP 01042024.pdf")
 text = []
 for p in np.arange(0, len(reader.pages), 1):
@@ -65,7 +67,7 @@ index = faiss.IndexFlatL2(d)
 index.add(text_embeddings)
 
 #index = faiss.read_index("./resourse/embeddings_ngap.faiss")
-
+"""
 print("Finish the model init process")
 
 def format_chat_prompt(message, chat_history):
@@ -85,6 +87,8 @@ context = [
     }
  ]
 
+max_new_tokens = 2048
+
 def respond(message, chat_history):
 
 
@@ -93,11 +97,13 @@ def respond(message, chat_history):
 
         context.append({'role':'user', 'content':f"{prompt}"})
 
-        tokenized_chat = tokenizer.apply_chat_template(context, tokenize=True, add_generation_prompt=True, return_tensors="pt")
+        #tokenized_chat = tokenizer.apply_chat_template(context, tokenize=True, add_generation_prompt=True, return_tensors="pt")
+
+        #outputs = model.generate(tokenized_chat, max_new_tokens=1000, temperature = 0.0)
 
-        outputs = model.generate(tokenized_chat, max_new_tokens=1000, temperature = 0.0)
+        #bot_message = tokenizer.decode(outputs[0]).split("<|assistant|>")[-1].replace("</s>","")
 
-        bot_message = tokenizer.decode(outputs[0]).split("<|assistant|>")[-1].replace("</s>","")
+        bot_message = model.generate(prompt=prompt, temp=0.5, top_k = 40, top_p = 1, max_tokens = max_new_tokens, streaming=False)
 
         context.append({'role':'assistant', 'content':f"{bot_message}"})