Searched defs:dscale_big (Results 1 – 2 of 2) sorted by relevance
101 float dscale_big = quant_max - zero_point; in _fake_quantize_grad_learnable_tensor_kernel_cuda() local
2598 float dscale_big = quant_max - zero_point; local2736 float dscale_big = quant_max - (*zero_point_input); local