Starting from the beginning

[Antonio Paes]

hello guys, I come up here because I am really stuck with caffe, already read the documentation did the test using the MNIST (which was the one who got), but still did not understand what the best language to use with caffe and how to use the network after trained (lenet_iter_10000 file), you just do me indicates a good tutorial of how to actually start would be great, despite official tutorials of caffe be good, I feel it lacks a middle thing "how to" even for beginners, and if someone want to ride a blog doing a walkthrough of how to install and use I'm disposition to help develop some examples.

Thank you.

[Nikiforos Pittaras]

What do you want to use the network for?

If you want to test it on mnist, you can run caffe test [...] to score the network on the mnist dataset (that's using the compiled C++ binary).

If you want to use it on a different dataset with different classes, you can finetune the network using caffe train -weights=<model_binary_name>.caffemodel [...], and test it on that dataset using the above.

If you want to use it as a generic feature extractor, you can use the extract_features C++ tool.

I am guessing you can also do all these with the python and matlab wrappers, so which language you use is purely a personal choice.