Chenglong Chen comments

Repositories
Issues
Comments

Results 3 comments of


                                            Chenglong Chen

有没有实现sparse输入的deepfm

这份代码写的比较早，没有使用 sparse tensor 的方式来做。我在另外一个 project XNN 有试过 sparse tensor & tf.nn.embedding_lookup_sparse，效率有些问题，可能是我实现的不对。

Why predict same test set, but given different predict_prob?

Are you using GPU? Can you check whether this happen again switching to CPU?

How can I build the "compute_image_mean.exe" for lmdb?

At the moment when I built this windows-version, I noticed there are some problems with the lmdb in windows, as mentioned here: https://initialneil.wordpress.com/2015/01/11/build-caffe-in-windows-with-visual-studio-2013-cuda-6-5-opencv-2-4-9/ So I haven't put much effort to...