Update model.py

model.py

@@ -104,7 +104,7 @@ def get_finance_llama_pipeline():
 
 	# Configure llama.cpp loader
 	threads = int(os.getenv("LLAMA_THREADS", str(os.cpu_count() or 2)))
-	ctx = int(os.getenv("LLAMA_CTX", "4096")) #Mặc định 2048 -> 1024
+	ctx = int(os.getenv("LLAMA_CTX", "131072")) #Mặc định 2048 -> 1024
 	batch = int(os.getenv("LLAMA_BATCH", "64")) #Mặc định 128 -> 64
 	use_mmap = os.getenv("LLAMA_MMAP", "1") in {"1","true","True","yes"}
 	use_mlock = os.getenv("LLAMA_MLOCK", "0") in {"1","true","True","yes"}