This contribution reports the results of the cross-check of JVET-AP0234. In NNVC-16.1 and ECM-19.1, in the Neural Network-based Intra Prediction (NNIP) implementation, layer activations use 16-bit signed integer, matrices of weights use 16-bit signed integer, and accumulators inside vector-matrix multiplications use 32-bit signed integer. JVET-AP0234 suggests to reduce the activation bit-width to 10-bits and the weight bit-width to 8 bits in order to avoid implementation-dependent behavior due to overflow and reduce hardware area in NNIP implementation.