fix: replace numpy.where with paddle.where for tensor operations in P…

[danghoangnhan]

Replaced np.where(temp.cpu().numpy() == True) with paddle.where(temp == True) to maintain computation on the PaddlePaddle device (GPU/CPU) without converting to NumPy. This avoids unnecessary data transfer between devices and keeps the operation within Paddle’s tensor graph.

[paddle-bot[bot]]

Thanks for your contribution!

[CLAassistant]

All committers have signed the CLA.