szha commented on issue #7999: Is ndarray api designed for users construct networks to predict, and symbol for training? URL: https://github.com/apache/incubator-mxnet/issues/7999#issuecomment-331676651 Symbolic way is usually faster. Most of the examples were written in symbolic way. For NDArray, you can use autograd to train in imperative way. https://mxnet.incubator.apache.org/versions/master/api/python/autograd.html?highlight=autograd#record-vs-pause Yet another alternative is to use the new interface Gluon, to use imperative way for development and debugging and to use symbolic way for large-scale training. There are many step-by-step tutorials on this in http://gluon.mxnet.io/ and https://zh.gluon.ai/. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
With regards, Apache Git Services