broliang / HRGAT

The code for hyper-node relational graph attention network
17 stars 2 forks source link

注释掉的代码 #5

Open lyriccs opened 4 months ago

lyriccs commented 4 months ago

请问这段代码为什么注释掉了,我把它加上之后一直报错

if self.use_text:

    #     edge_data_t = self.comp(edges.src['t'], self.rel[edge_type])  # [E, in_channel]
    #     msg_t = torch.cat([torch.matmul(edge_data_t[:edge_num // 2, :], self.in_w_t),
    #                        torch.matmul(edge_data_t[edge_num // 2:, :], self.out_w_t)])
    #     msg_t = torch.cat((Variable(torch.ones(msg_t.shape[0], 1).type(self.DTYPE).to(self.device), requires_grad=False), msg_t), dim=1)
    #     fused_msg_t = torch.matmul(msg_t, self.text_facotr)
    #
    # if self.use_img:
    #     edge_data_i = self.comp(edges.src['i'], self.rel[edge_type])  # [E, in_channel]
    #     msg_i = torch.cat([torch.matmul(edge_data_i[:edge_num // 2, :], self.in_w_i),
    #                        torch.matmul(edge_data_i[edge_num // 2:, :], self.out_w_i)])
    #     msg_i = torch.cat((Variable(torch.ones(msg_i.shape[0], 1).type(self.DTYPE).to(self.device), requires_grad=False), msg_i), dim=1)
    #     fused_msg_i = torch.matmul(msg_i, self.img_facotr)
    #
    #     # msg = torch.cat([msg, msg_i], dim=1)
ttzhang0417 commented 4 months ago

您好,请问你解决这个问题了吗,我也有同样的疑问

lyriccs commented 4 months ago

我正在改这段代码

---原始邮件--- 发件人: @.> 发送时间: 2024年5月3日(周五) 晚上7:07 收件人: @.>; 抄送: @.**@.>; 主题: Re: [broliang/HRGAT] 注释掉的代码 (Issue #5)

您好,请问你解决这个问题了吗,我也有同样的疑问

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>