Searched refs:bias_scale (Results 1 – 2 of 2) sorted by relevance
261 const double bias_scale = biasShape.scale; in GetQuantizedConvolutionMultipler() local264 NN_OPS_CHECK(std::abs(input_product_scale - bias_scale) <= in GetQuantizedConvolutionMultipler()265 1e-6 * std::min(input_product_scale, bias_scale)); in GetQuantizedConvolutionMultipler()
1132 * with zeroPoint of 0 and bias_scale == input_scale * filter_scale.4313 * of 0 and bias_scale == input_scale * filter_scale. For filter tensor4316 * 0 and bias_scale of 0. The actual scale of each value 'i' is equal to4317 * bias_scale[i] = input_scale * filter_scale[i].4358 * of 0 and bias_scale == input_scale * filter_scale. For filter tensor4361 * 0 and bias_scale of 0. The actual scale of each value 'i' is equal to4362 * bias_scale[i] = input_scale * filter_scale[i].5748 * with zeroPoint of 0 and bias_scale == input_scale * filter_scale.5751 * and bias_scale of 0. The actual scale of each value 'i' is equal to5752 * bias_scale[i] = input_scale * filter_scale[i].[all …]