Expand description
Convolution layer operations for quantized neural networks.
Functionsยง
- convolve_
1_ x_ n_ s8 - Performs 1D Temporal Convolution for int8 tensors (
convolve_1_x_n_s8). - convolve_
per_ channel_ s8 - Performs standard 2D Convolution with per-channel quantization.
- convolve_
s8 - Performs standard 2D Convolution with per-tensor quantization.
- depthwise_
conv_ per_ channel_ s8 - Performs Depthwise 2D Convolution with per-channel quantization.
- transpose_
conv_ s8 - Performs Transposed 2D Convolution (Deconvolution) for int8 tensors with per-channel quantization.