Open Orca, an open sourced replication of Microsofts Orca is in development! Heres the dataset!

Today we are releasing a dataset that lets open source models learn to think like GPT-4!

We call this Open Orca, as a tribute to the team who has released the Orca paper describing the data collection methods we have attempted to replicate in an open-source manner for the benefit of humanity.

With this data, we expect new open source models to be developed which are smaller, faster, and smarter than ever before because were going to be the ones …

call collection data data collection dataset deeplearning development gpt gpt-4 learn open source paper replicate replication team think