mnli_example实验相关 about textbrewer

[airaria]

并没有使用transformer，用的是其原先版本pytorch_pretrained_bert，也在mnli_example目录下。 可能是权重问题， 你把assert len(missing_keys)==0 改成 assert len(missing_keys)==0,missing_keys，看下到底是什么权重没匹配上 在 2021-02-03 16:57:43，"qq260612718" <[email protected]> 写道： 您好，请问mnli_example这个实验环境是多少？如：pytorch和transformers等。 在多次尝试运行时都会出现以下问题： 2021/02/03 16:54:38 - INFO - utils - Loading features from cached file data_root_dir/SST2/cached_train_128_sst-2 2021/02/03 16:54:39 - INFO - utils - Loading features from cached file data_root_dir/SST2/cached_dev_128_sst-2 2021/02/03 16:54:39 - INFO - Main - Data loaded Traceback (most recent call last): File "main.trainer.py", line 220, in main() File "main.trainer.py", line 160, in main assert len(missing_keys)==0 AssertionError — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

Closing the issue, since no updates observed. Feel free to re-open if you need any further assistance.

[qq260612718]

谢谢，原始实验模型已经可以运行了。我想将您该mnli_example实验所用到的BERT-Base, Cased: 12-layer, 768-hidden, 12-heads , 110M parameters,替换成BERT-Base, Multilingual Cased (New, recommended): 104 languages, 12-layer, 768-hidden, 12-heads, 110M parameters尝试运行，但所有权重似乎都不能匹配上，我检查了两个模型bert_config基本相同，请问该预训练模型可以修改替換吗？

[airaria]

如果匹配不上，你print一下bert-base模型和multilingual bert模型的权重(state_dict)的key，它们应该是一样的，看看是不是这里不同？ 在 2021-02-18 17:37:56，"qq260612718" <[email protected]> 写道： 谢谢，原始实验模型已经可以运行了。我想将您该mnli_example实验所用到的BERT-Base, Cased: 12-layer, 768-hidden, 12-heads , 110M parameters,替换成BERT-Base, Multilingual Cased (New, recommended): 104 languages, 12-layer, 768-hidden, 12-heads, 110M parameters尝试运行，但所有权重似乎都不能匹配上，我检查了两个模型bert_config基本相同，请问该预训练模型可以修改替換吗？ — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[Soulscb]

谢谢，原始实验模型已经可以运行了。我想将您该mnli_example实验所用到的BERT-Base, Cased: 12-layer, 768-hidden, 12-heads , 110M parameters,替换成BERT-Base, Multilingual Cased (New, recommended): 104 languages, 12-layer, 768-hidden, 12-heads, 110M parameters尝试运行，但所有权重似乎都不能匹配上，我检查了两个模型bert_config基本相同，请问该预训练模型可以修改替換吗？

请问你是怎么运行起来的呀？我最近也遇到这个问题了，尝试了各种方法都没运行起来