This is the primary idea of Caffe2 API:
use Python to conveniently compose nets to train your model,
pass those nets to C++ code as serialized protobuffers,
and then let the C++ code run the nets with full performance.
However, after I compose my net, export my net to protobuf, I don't know how to train my net will full performance by C++, I had dived for a while, trying to find a tutorial on the website or an example in caffe2's repository, but maybe I missed something, I can not find it.
In caffe2's binaries folder, where is a run_play.cc, I guess it might be where to start. but I am not sure, and I have no idea about how to use it.
Can anyone so kind as to tell me what I have missed? thank you very much.
yxchng
commented
7 years ago
Hi all, I am new to caffe2 (have used caffe for a while).
In the intro-tutorial, there is paragraph said:
However, after I compose my net, export my net to protobuf, I don't know how to train my net will full performance by C++, I had dived for a while, trying to find a tutorial on the website or an example in caffe2's repository, but maybe I missed something, I can not find it.
In caffe2's binaries folder, where is a run_play.cc, I guess it might be where to start. but I am not sure, and I have no idea about how to use it.
Can anyone so kind as to tell me what I have missed? thank you very much.