Switch to open model

app.py

@@ -6,7 +6,6 @@ import torch
 import numpy as np
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
-
 # Load corpus
 print("Loading dataset...")
 dataset = load_dataset("rag-datasets/rag-mini-wikipedia", "text-corpus")
@@ -26,8 +25,8 @@ index.add(corpus_embeddings_np)
 reranker = CrossEncoder("cross-encoder/ms-marco-MiniLM-L-6-v2")
 
 # Generator (choose one: local HF model or OpenAI)
-tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.3")
-model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.3", device_map="auto", torch_dtype=torch.float16)
+tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-4")
+model = AutoModelForCausalLM.from_pretrained("microsoft/phi-4", device_map="auto", torch_dtype=torch.float16)
 generator = pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=150)
 
 def rag_pipeline(query):