Full threadjarrell_mark·Deep Learning could in theory run on this using Intel PlaidML's Keras implementation in OpenCL https://github.com/plaidml/plaidmlView on HN