| 207 | return x1 * x2 |
| 208 | |
| 209 | class LayerNormFunction(Function): |
| 210 | |
| 211 | @staticmethod |
| 212 | def forward(ctx, x, weight, bias, eps): |
| 213 | ctx.eps = eps |
| 214 | N, C, H, W = x.size() |
| 215 | mu = x.mean(1, keepdim=True) |
| 216 | var = (x - mu).pow(2).mean(1, keepdim=True) |
| 217 | y = (x - mu) / (var + eps).sqrt() |
| 218 | ctx.save_for_backward(y, var, weight) |
| 219 | y = weight.view(1, C, 1, 1) * y + bias.view(1, C, 1, 1) |
| 220 | return y |
| 221 | |
| 222 | @staticmethod |
| 223 | def backward(ctx, grad_output): |
| 224 | eps = ctx.eps |
| 225 | |
| 226 | N, C, H, W = grad_output.size() |
| 227 | y, var, weight = ctx.saved_variables |
| 228 | g = grad_output * weight.view(1, C, 1, 1) |
| 229 | mean_g = g.mean(dim=1, keepdim=True) |
| 230 | |
| 231 | mean_gy = (g * y).mean(dim=1, keepdim=True) |
| 232 | gx = 1. / sqrt(var + eps) * (g - y * mean_gy - mean_g) |
| 233 | return gx, (grad_output * y).sum(dim=3).sum(dim=2).sum(dim=0), grad_output.sum(dim=3).sum(dim=2).sum( |
| 234 | dim=0), None |
| 235 | |
| 236 | class LayerNorm2d(nn.Module): |
| 237 |
nothing calls this directly
no outgoing calls
no test coverage detected