A naukri.com initiative
Cloudblog
2M
431
Image Credit: Cloudblog
BigQuery ML is now compatible with open-source gen AI models
- BigQuery Machine Learning allows you to use large language models (LLMs) like Gemini for tasks such as entity extraction, sentiment analysis, and more using SQL syntax.
- Support for open-source LLMs from Vertex AI Model Garden, including models from Hugging Face, is now available, providing developers with a wider model selection.
- Integration example using the Meta Llama 3.3 70B model demonstrates the process, but any of over 170K text generation models from Hugging Face can be used following similar steps.
- Steps involve hosting the model on a Vertex endpoint and creating a remote model in BigQuery to allow inference.
- Performing inference involves extracting structured data from unstructured transcripts, as shown with a medical transcripts dataset example.
- Create a table in BigQuery to analyze the data, and use the Llama model for data extraction with SQL statements.
- Analytics can be performed on the results to gain insights, such as identifying common diseases in specific demographics.
- Utilize the output from the Llama endpoint for further analysis and querying, such as identifying common diseases in females above 30 years of age.
- Get started today with BigQuery ML and the Vertex Model Garden integration to leverage open models or tuned/distilled models for your analytics needs.
- Learn more about using BigQuery with different models in the official documentation.
Read Full Article
25 Likes
For uninterrupted reading, download the app