Skip to content

Instantly share code, notes, and snippets.

@dennyglee

dennyglee/llama-index_starter_dbrx.py

Last active April 10, 2024 07:25
Show Gist options
  • Save dennyglee/3a4fd9042c283549b727e081397842da to your computer and use it in GitHub Desktop.
Save dennyglee/3a4fd9042c283549b727e081397842da to your computer and use it in GitHub Desktop.
Download ZIP
Llama_Index Starter Example using DBRX
Raw
llama-index_starter_dbrx.py
#
# LlamaIndex Starter Example with DBRX
#
# Based on the LlamaIndex Starter Example
# https://docs.llamaindex.ai/en/stable/getting_started/starter_example/
#
# Ensure you have installed both llama_index and Databricks integration
# e.g., pip install llama_index llama-index-llms-databricks
#
# Ensure you have configured the following environment variables
# export DATABRICKS_TOKEN=...
# export DATABRICKS_SERVING_ENDPOINT=...
#
# Load data
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings
documents = SimpleDirectoryReader("data").load_data()
# Configure to use DBRX
from llama_index.llms.databricks import Databricks
# define LLM
Settings.llm = Databricks(model="databricks-dbrx-instruct")
# Define index
index = VectorStoreIndex.from_documents(documents)
# Query your data
my_question = "What did the author do growing up?"
query_engine = index.as_query_engine()
response = query_engine.query(my_question)
print("===========================================")
print("Question: %s \n" % my_question)
print("DBRX Response: %s" % response)
print("===========================================")
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
  翻译：