how could I transform between NumSharp.NDArray with Tensorflow.Numpy.NDArray? #465

Activity

1. Oceania2018 commented on Nov 27, 2021

   

   Oceania2018

   on Nov 27, 2021

   Member

   More actions

   What are the missing methods?
2. Nucs commented on Sep 7, 2026

   

   Nucs

   on Sep 7, 2026

   Member

   More actions

   @cross-hello, there's no direct type converter between the two libraries, but both wrap raw buffers, so you can bridge in two ways (0.70.0, #628):

   1. Via the byte buffer - tobytes out, np.frombuffer in:

   var a = np.array(new[] { 1.0, 2, 3 });
   byte[] raw = a.tobytes();
   
   var back = np.frombuffer(raw, "float64");   // [1. 2. 3.]
   // hand `raw` to Tensorflow.NumPy the same way, or read its buffer back through frombuffer

   2. Via NumSharp.Interop.pythonnet - cross through a real numpy array (arr.ToNumpy() / pyObj.AsNDArray()), which any numpy-based stack accepts.

   Closing.
3. Nucs

   closed this as completedon Sep 7, 2026