autograd_Sigmoid.html

<!DOCTYPE html>
<html style="height: 100%">
<head>
    <meta charset="utf-8">
    <script type="text/javascript" src="torch.js"></script>
</head>
<body style="height: 100%; margin: 0">
        <script type="text/javascript">
            function sigmoid(x) {
                return torch.function(torch.const(1).div(torch.const(1).add(torch.const(Math.E).pow(torch.tensor(x).minus()))));
            }

            function activation(x) {
                return 1.0 / (1.0 + Math.exp(-x));
            }

            function derive(x) {
                return activation(x) * (1-activation(x));
            }

            // 自动求导
            var y_ = sigmoid(torch.tensor(10));
            var der = y_.head.derive();
            console.log(der);

            // 解析法求导
            var y_ = activation(10);
            var der = derive(10);
            console.log(der);
        </script>
</body>
</html>