orca dataset

StableBeluga1-Delta is a Llama65B model fine-tuned on an Orca-style dataset, designed for text generation tasks. It requires applying delta weights to the base LLaMA 65B model to obtain the full Stable Beluga 1 model. The model is licensed under the Non-Commercial Creative Commons license (CC BY-NC-4.0) and is intended for developers who need a powerful language model for generating text based on instructions. It is particularly useful for tasks that require following complex guidelines or creating content based on specific inputs. However, it is important to note that the model may produce inaccurate or biased outputs, and developers should perform safety testing before deployment.