tf.raw_ops.FakeQuantWithMinMaxVarsPerChannelGradient

Compute gradients for a FakeQuantWithMinMaxVarsPerChannel operation.

gradientsA Tensor of type float32. Backpropagated gradients above the FakeQuantWithMinMaxVars operation, shape one of: [d], [b, d], [b, h, w, d].
inputsA Tensor of type float32. Values passed as inputs to the FakeQuantWithMinMaxVars operation, shape same as gradients. min, max: Quantization interval, floats of shape [d].
minA Tensor of type float32.
maxA Tensor of type float32.
num_bitsAn optional int. Defaults to 8. The bitwidth of the quantization; between 2 and 16, inclusive.
narrow_rangeAn optional bool. Defaults to False. Whether to quantize into 2^num_bits - 1 distinct values.
nameA name for the operation (optional).

A tuple of Tensor objects (backprops_wrt_input, backprop_wrt_min, backprop_wrt_max).
backprops_wrt_inputA Tensor of type float32.
backprop_wrt_minA Tensor of type float32.
backprop_wrt_maxA Tensor of type float32.