Closed lzf00 closed 2 years ago
RuntimeError: mat1 and mat2 shapes cannot be multiplied (1224x768 and 1024x256),而且每次运行,这个mat1的矩阵维度会变。。。。。不知道在哪里修改,是设置了随机值吗
怎么会出现768,你用的是bert-base吗? 维度每次都变,肯定是random sample的原因,每次mini-batch的长度不会一致
我也是这个问题,请问作者和楼主解决了吗
我也是这个问题,请问作者和楼主解决了吗
您好,我也出现了这个,试了autotokenizer也没用,请问您解决了吗
没有解决,不过你是从哪里知道我邮箱的啊,我这个问题问了很久了,在哪儿问的我都忘记了
------------------ 原始邮件 ------------------ 发件人: "xhw205/GPLinker_torch" @.>; 发送时间: 2023年3月25日(星期六) 晚上6:34 @.>; @.**@.>; 主题: Re: [xhw205/GPLinker_torch] 模型报错了。。RuntimeError: mat1 and mat2 shapes cannot be multiplied (1368x768 and 1024x256) (Issue #7)
我也是这个问题,请问作者和楼主解决了吗
您好,我也出现了这个,试了autotokenizer也没用,请问您解决了吗
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>
main.py中 RawGlobalPointer(hiddensize=1024)这些地方对应改成base的hidden size了吗?
Traceback (most recent call last): File "/content/drive/MyDrive/GPLinker_torch-main/main.py", line 85, in
logits1, logits2, logits3 = net(batch_token_ids, batch_mask_ids, batch_token_type_ids)
File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, kwargs)
File "/content/drive/MyDrive/GPLinker_torch-main/main.py", line 53, in forward
mention_outputs = self.mention_detect(outputs, batch_mask_ids)
File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, *kwargs)
File "/content/drive/MyDrive/GPLinker_torch-main/nets/gpNet.py", line 67, in forward
outputs = self.dense(last_hidden_state)
File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(input, kwargs)
File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/linear.py", line 103, in forward
return F.linear(input, self.weight, self.bias)
RuntimeError: mat1 and mat2 shapes cannot be multiplied (1368x768 and 1024x256)