Efficient continual pre-training LLMs for financial domains
AWS Machine Learning
MARCH 28, 2024
For example, the training data used for BloombergGPT is 51% domain-specific documents, including financial news, filings, and other financial materials. Some of the content is based on the paper Efficient Continual Pre-training for Building Domain Specific Large Language Models. This creates a large number of documents over the years.
Let's personalize your content