2020-09-15

#nn.Linear的操作是torch.mm(h_t, torch.transpose(weight,1,0))
layer = nn.Linear(
            in_features=4, out_features=4,bias = False
        )
transformed_h_t = layer(h_t)
params = list(layer.named_parameters())
print(params[0])
print("transformed_h_t",transformed_h_t)
weight = torch.tensor([[ 0.4935, -0.2656,  0.4077, -0.4063],
        [-0.1951,  0.2847, -0.4069, -0.0827],
        [-0.4420,  0.1278,  0.2191, -0.3036],
        [-0.3300, -0.2710,  0.2737,  0.2945]])
# bias = torch.tensor([-0.1908, -0.1736,  0.2886,  0.0754])
print(torch.transpose(weight,1,0))
h_t_mm = torch.mm(h_t, torch.transpose(weight,1,0))
print("h_t_mm",h_t_mm)
  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值