Minor data type modification

[darkthecross]

No description provided.

[HuanWeng]

Sorry I don't know why it is necessary. Could you show me some example showing its importance? Many thanks!

[darkthecross]

It is not really necessary I guess. When I was using this function, I somehow got a matrix with first three rows with type np.float32 and the last row with type np.object, which made me uncomfortable. But I guess numpy is able to handle following calculation correctly in this case.

On Sun, Mar 3, 2019 at 4:31 PM HuanWeng ***@***.***> wrote: Sorry I don't know why it is necessary. Could you show me some example showing its importance? Many thanks! — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#19 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AB0U3Kvk2AtfsA_09EcqJNQ1OyaJ7yBxks5vTGldgaJpZM4baTHR> .

-- 樊明祥 清华大学微纳电子系 IME, Tsinghua University TEL: 13521052496