search

IAAR-Shanghai / CRUD_RAG

CRUD-RAG: A Comprehensive Chinese Benchmark for Retrieval-Augmented Generation of Large Language Models
https://arxiv.org/abs/2401.17043
233 stars 20 forks source link

llama_index and langchain version issue #14

Closed XinyuYe-Intel closed 3 months ago

XinyuYe-Intel commented 3 months ago

Hi, could you please share the version of llama_index and langchain that you used? I've tried latest version, but seems llama_index newer than 0.9.48 would be incompatible with your code, so I used llama_index 0.9.36, which appears compatible, but langchain version issue still exists, seems caused by HuggingfaceEmbeddings inherited from langchain.schema.embeddings.Embeddings does not have get_agg_embedding_from_queries function. image

Below is the command I used:

python quick_start.py     --temperature 0.1   --max_new_tokens 1280     --shuffle True     --docs_type 'txt'   --docs_path data/80000_docs   --chunk_size 128   --chunk_overlap 0   --retriever_name 'base'   --collection_name milvus_test  --retrieve_top_k 8   --task 'all'   --num_threads 20   --show_progress_bar True   --construct_index
haruhi-sudo commented 3 months ago

Hello, I have updated the requirement.txt file.

XinyuYe-Intel commented 3 months ago

Thanks! It works now.