search

Mikolaj / horde-ad

Higher Order Reverse Derivatives Efficiently - Automatic Differentiation library based on the paper "Provably correct, asymptotically efficient, higher-order reverse-mode automatic differentiation"
BSD 3-Clause "New" or "Revised" License
34 stars 6 forks source link

Vectorize and AD code that transforms tensors of Double and tensors of Float #97

Closed Mikolaj closed 1 year ago

Mikolaj commented 1 year ago

and of CInt or Int64 and of other types.

This is postponed, because this is hard. See #91 for how even permitting tensors of integers, in addition to tensors of the chosen float, is hard.

Mikolaj commented 1 year ago

This is now done (tcast and scast and tfromIntegral and sfromIntegral). Closing.