Can it be used with torch's dataparallel to train model on multi-gpus?

[poisson-AI]

I want to know how I can use it with dataparallel. What should I change in my code when I build my model on several gpus?

[LemonMi]

Have you solved this problem? I meet the same issue... i'm wondering If you could tell me how you deal with it?

[poisson-AI]

兄dei，好像是同胞？这个问题我当时解决掉了，不知道你解决掉了没有，如果需要的话有空我给你找找

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "LemonMi"<[email protected]&gt;; 发送时间: 2019年12月31日(星期二) 下午5:59 收件人: "nearai/torchfold"<[email protected]&gt;; 抄送: "10671469"<[email protected]&gt;; "Author"<[email protected]&gt;; 主题: Re: [nearai/torchfold] Can it be used with torch&#39;s dataparallel to train model on multi-gpus? (#7) Have you solved this problem? I meet the same issue... i'm wondering If you could tell me how you deal with it? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[LemonMi]

@guobaisong 哈哈是同胞。。 我目前是把所有用到的网络都放到nn.DataParallel()里了，运行倒是没问题，就是不知道对不对。如果方便的话，可能得麻烦你有空找一下，非常感谢

[poisson-AI]

就是可以直接这样用的，没有问题～我当时应该也是这样搞的

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "LemonMi"<[email protected]&gt;; 发送时间: 2020年1月5日(星期天) 下午5:59 收件人: "nearai/torchfold"<[email protected]&gt;; 抄送: "10671469"<[email protected]&gt;; "Mention"<[email protected]&gt;; 主题: Re: [nearai/torchfold] Can it be used with torch&#39;s dataparallel to train model on multi-gpus? (#7) @guobaisong 哈哈是同胞。。 我目前是把所有用到的网络都放到nn.DataParallel()里了，运行倒是没问题，就是不知道对不对。如果方便的话，可能得麻烦你有空找一下，非常感谢 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[LemonMi]

@guobaisong 好的好的～谢谢

[poisson-AI]

哈哈哈没事没事，当时我找了好久怎么train不同形状的树

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "LemonMi"<[email protected]&gt;; 发送时间: 2020年1月5日(星期天) 晚上6:04 收件人: "nearai/torchfold"<[email protected]&gt;; 抄送: "10671469"<[email protected]&gt;; "Mention"<[email protected]&gt;; 主题: Re: [nearai/torchfold] Can it be used with torch&#39;s dataparallel to train model on multi-gpus? (#7) @guobaisong 好的好的～谢谢 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.