debug

app.py

@@ -15,7 +15,7 @@ from transformers import (
 import torch.nn.functional as F
 from collections import defaultdict
 HF_TOKEN = os.getenv("HF_token")
-CHAT_MODEL_ID = "NousResearch/Meta-Llama-3-8B-Instruct"
+CHAT_MODEL_ID = "meta-llama/Llama-3.1-8B-Instruct"
 EMB_MODEL_ID  = "mixedbread-ai/mxbai-embed-large-v1"
 MAX_PROMPT_TOKENS = 8192
 
@@ -190,7 +190,7 @@ def answer(system: str, context: str, question: str,
             top_k=top_k_tok
         )
         full   = tokenizer.decode(output[0], skip_special_tokens=True)
-        reply  = full.split("<|im_start|>assistant")[-1].strip()
+        reply  = full.split("<|im_start|>assistant")[-1].strip() + prompt
         return reply
           
     except Exception as e: