jiaohui
jiaohui
i'v tried several times and found that this configuration works for me: python=3.8 pytorch==1.10.0 cuda=11.1 fairseq==0.10.0 gpu=3090
> try to add `self.args = args` in the def **init**(self, args, src_dict, tgt_dict): of translation_from_pretrained_bart.py It works for me. thank you , it also works for me.
那我要如何区分相同tag的不同变量呀？
好的，试了下挺好的
好的我试试 ------------------ 原始邮件 ------------------ 发件人: ***@***.***>; 发送时间: 2022年9月17日(星期六) 下午5:21 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [PaddlePaddle/PaddleNLP] [Question]: 在aistudio运行run_moe_pretrain.py失败 (Issue #3287) 报错可以试下把paddlepaddle-gpu升级到最新版本 — Reply to this email directly, view it on GitHub,...
> 报错可以试下把paddlepaddle-gpu升级到最新版本 您好，paddle需要哪个版本呢？ 升级到最新版本报错了。
> 报错可以试下把paddlepaddle-gpu升级到最新版本 您好，我已经安装了nightly版本paddle，基本上可以运行了，只不过加载数据时候格式似乎不对，这是什么原因呢？ 复现流程： 1.安装moe_pretrain需要的依赖 pip install regex LAC 2.安装paddle nightly cuda11.2 python -m pip install paddlepaddle-gpu==0.0.0.post112 -f https://www.paddlepaddle.org.cn/whl/linux/gpu/develop.html 3.设置环境变量 若不设置会在安装paddlenlp后报错： importError: libpython3.7m.so.1.0: cannot open shared object file: No such...
> I also encountered it, basically, pretraining RoBERTa with whole-word-masking will fail because it assumes the existence of `args` field, which was replaced with hydra's `cfg`. > > A quick...
Nll_loss here is used to calculate the negative log likelihood - logP(D|wi) in the paper of blundell, In my opinion, here should not apply log_softmax ,calculate mean probs and then...
> copy from somwhere. `import zhon.hanzi or from zhon import hanzi` > copy from somwhere. `import zhon.hanzi or from zhon import hanzi` it works ! thank you!!!