📄️ Examples in DSPy
Working in DSPy involves training sets, development sets, and test sets. This is like traditional ML, but you usually need far fewer labels (or zero labels) to use DSPy effectively.
📄️ Utilizing Built-in Datasets
It's easy to use your own data in DSPy: a dataset is just a list of Example objects. Using DSPy well involves being able to find and re-purpose existing datasets for your own pipelines in new ways; DSPy makes this a particularly powerful strategy.
📄️ Creating a Custom Dataset
We've seen how to work with with Example objects and use the HotPotQA class to load the HuggingFace HotPotQA dataset as a list of Example objects. But in production, such structured datasets are rare. Instead, you'll find yourself working on a custom dataset and might question: how do I create my own dataset or what format should it be?