Closed wenruij closed 6 years ago
Hi wenruij, that is not a bug. We have some constraints on the format of input data. Basically, it's a field-wise format, each line should contain exactly FIELD_CNT fields, while each field can contain 1 or more features. I am thinking about that I need to provide a more detailed descriptive file to illustrate this data format.
@Leavingseason Thanks for your reply. Looking forward to your deep learning based factorization toolkit and all mentioned models.
@Leavingseason Seems there's a bug for function
_build_embedding
:You do a reshape after the
tf.nn.embedding_lookup_sparse
, let me imitate a simple example:We will get the following errors when running
embedding
with tf.Session: