OpenCodeGeneticInstruct: A large-scale dataset of coding instructions for improving the code generation capabilities of LLMs
Data Overview
OpenCodeGeneticInstruct comprises more than 15M coding instructions in python which is generated synthetically with the Genetic-Instruct [1] approach.
This dataset can be used for supervised fine-tuning (SFT) of LLMs to improve their code genearation capability.
Each sample includes a coding question/instruction and its corrsponding… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/OpenCodeGeneticInstruct.
