LLMs

Fine-tune CodeLlama with Gretel's Synthetic Text-to-SQL on BIRDBench, achieving a 36% relative improvement in EX and 38% in VES.

Gretel Navigator now enables fine-tuning for tabular datasets containing numeric, categorical, free text, JSON, and sequential (time series) data.

Gretel Navigator’s synthetic data generation outperformed OpenAI's GPT-4 by 25.6%, surpassed Llama3-70b by 48.1%, and exceeded human expert-curated data by 73.6%.

Gretel partners with Databricks to seamlessly integrate synthetic data workflows and improve model performance for Enterprise AI.

How to use Gretel Navigator for generating diverse, high-quality training data to create better language models.

Effective strategies for leveraging high-quality synthetic data to improve RAG model performance.

How differential privacy can generate provably private synthetic text data for a variety of enterprise AI applications.

How to automatically generate missing tabular data that maintains contextual relevance.

Automate hyperparameter sweeps to create the best synthetic data for your task 🧹