COPYCAT_RANK must be < COPYCAT_WORLD_SIZE.

COPYCAT_RANK: the current process rank relative to other processes.COPYCAT_MAIN_PORT: the main port for process 0.Process 0 runs on this IP address and only this process saves the contact sheets, graphs, and training model checkpoints. Note:Cop圜at can train using Nuke Render Licences on distributed machinesĭistributed Training in Cop圜at is controlled through the following environment variables: Note:You must also install Nuke on the worker machines for distributed training to operate as expected. The image shows a typical distributed training set up using a render manager to distribute the training load. We expect it would work using 15.0 but this has not been tested.Īs an artist, you work as normal on your workstation and the render manager handles the distribution of training tasks to the available workers. Note:Distributed Cop圜at training is qualified on Nuke 14.1 and 15.0.ĭistributed Cop圜at training via OpenCue has been validated internally using 14.1. Offloading the training onto different machines also means that you can continue working when Cop圜at is training models in the background. Share the Load with Distributed Cop圜at TrainingĬop圜at can share the training load across multiple machines, with a familiar interface, so that distributing training feels just like distributing rendering using standard render farm applications such as OpenCue.

Instaling NUKE Studio 14.1v1