[MLsys2026 Best Paper]: https://arxiv.org/abs/2506.08276. RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.
Python aifaissgpt-osslangchainllama-indexllmlocalstorageoffline-firstollamaprivacypythonragretrieval-augmented-generationvector-databasevector-searchvectors
fix some hf problem
Fangzhou6688ca09440dda
diff --git a/examples/google_history_reader_leann.py b/examples/google_history_reader_leann.pyindex f8acf90..4e92f7d 100644--- a/examples/google_history_reader_leann.py+++ b/examples/google_history_reader_leann.py@@ -1,12 +1,17 @@import osimport asyncio-import dotenv+try:+ import dotenv+ dotenv.load_dotenv()+except ModuleNotFoundError:+ # python-dotenv is not installed; skip loading environment variables+ dotenv = Nonefrom pathlib import Pathfrom typing import List, Anyfrom leann.api import LeannBuilder, LeannSearcher, LeannChatfrom llama_index.core.node_parser import SentenceSplitter-dotenv.load_dotenv()+# dotenv.load_dotenv() # handled above if python-dotenv is availabledef create_leann_index_from_multiple_chrome_profiles(profile_dirs: List[Path], index_path: str = "chrome_history_index.leann", max_count: int = -1):"""@@ -190,7 +195,7 @@ async def query_leann_index(index_path: str, query: str):query: The query string"""print(f"\n[PHASE 2] Starting Leann chat session...")- chat = LeannChat(index_path=index_path)+ chat = LeannChat(index_path=index_path, llm_config={"type": "hf", "model": "Qwen/Qwen3-0.6B"})print(f"You: {query}")chat_response = chat.ask(@@ -227,7 +232,7 @@ async def main():return# Create or load the LEANN index from all sources- index_path = create_leann_index_from_multiple_chrome_profiles(profile_dirs, INDEX_PATH)+ index_path = create_leann_index_from_multiple_chrome_profiles(profile_dirs, INDEX_PATH,1000)if index_path:# Example queriesdiff --git a/packages/leann-core/src/leann/chat.py b/packages/leann-core/src/leann/chat.pyindex 47b6e41..62d68a5 100644--- a/packages/leann-core/src/leann/chat.py+++ b/packages/leann-core/src/leann/chat.py@@ -128,6 +128,19 @@ class HFChat(LLMInterface):self.pipeline = pipeline("text-generation", model=model_name, device=device)def ask(self, prompt: str, **kwargs) -> str:+ # Map OpenAI-style arguments to Hugging Face equivalents+ if "max_tokens" in kwargs:+ # Prefer user-provided max_new_tokens if both are present+ kwargs.setdefault("max_new_tokens", kwargs["max_tokens"])+ # Remove the unsupported key to avoid errors in Transformers+ kwargs.pop("max_tokens")++ # Handle temperature=0 edge-case for greedy decoding+ if "temperature" in kwargs and kwargs["temperature"] == 0.0:+ # Remove unsupported zero temperature and use deterministic generation+ kwargs.pop("temperature")+ kwargs.setdefault("do_sample", False)+# Sensible defaults for text generationparams = {"max_length": 500,