I don’t still understand that the Features (inputs) must be in row or column format…
There is no universal standard on this. You may find datasets that use either orientation.
But Tensoflow accepts which one?
20 Haz 2025 Cum, saat 23:35 tarihinde Tom Mosher via DeepLearning.AI <notifications@dlai.discoursemail.com> şunu yazdı:
In TensorFlow, the first dimension (index 0) is the “samples” dimension. The other dimensions are the “features” dimensions.
As Tom says, you may find other environments in which the “samples” dimension is not the first dimension. If you are given a training dataset and are planning to use TF to build a model to process that data, then you’ll need to make sure you understand the structure of the data and rearrange it if necessary to get the “samples” dimension first before you feed that data to your TF model.