callback变红

[lili-li-cpu]

提问时请尽可能提供如下信息：

基本信息

• 你使用的操作系统: win11
• 你使用的Python版本:
• 你使用的Pytorch版本: 1.8.2+cu111
• 你使用的bert4torch版本: 0.2.8
• 你加载的预训练模型:chinese-bert-wwm-ext

核心代码

# 请在此处贴上你的核心代码

输出信息

# 请在此处贴上你的调试输出
<img width="768" alt="image" src="https://github.com/Tongjilibo/bert4torch/assets/66951248/0d006ca0-63af-4b4f-800e-1bc7680447fa">


### 自我尝试
1、首先，我的代码出现了这个问题
raceback (most recent call last):
File "train.py", line 423, in
model.fit(train_dataloader, steps_per_epoch=None, epochs=epochs, callbacks=[evaluator])
File "/usr/local/lib/python3.8/dist-packages/torch4keras/model.py", line 225, in fit
self.output, self.loss, self.loss_detail = self.train_step(self.train_X, self.train_y)
File "/usr/local/lib/python3.8/dist-packages/torch4keras/model.py", line 111, in train_step
output = self._forward(train_X)
File "/usr/local/lib/python3.8/dist-packages/torch4keras/model.py", line 94, in _forward
return self.get_module().forward(*inputs, **input_kwargs)
TypeError: forward() takes 2 positional arguments but 4 were given
2、自我尝试
我在您的issue里面找到了类似问题，然后按照您说的
<img width="590" alt="image" src="https://github.com/Tongjilibo/bert4torch/assets/66951248/a2dfcb43-a201-4ec8-9347-d00827f5195e">
3、在变完版本后，出现了这个问题
<img width="454" alt="image" src="https://github.com/Tongjilibo/bert4torch/assets/66951248/4aaa5709-4388-437a-b92d-ef9fd19130c3">
Callback标红
<img width="774" alt="image" src="https://github.com/Tongjilibo/bert4torch/assets/66951248/4f436367-0b55-41e0-9fda-08a8820c2342">
谢谢，，这就是我的问题

[lili-li-cpu]

Traceback (most recent call last):
File "D:/PycharmProjects/9.26-guan/project/chushi.py", line 11, in
from bert4torch.snippets import sequence_padding, Callback, ListDataset, seed_everything
ImportError: cannot import name 'Callback' from 'bert4torch.snippets' (D:\PycharmProjects\9.26-guan\bert4torch\snippets.py)
这就是当前的问题

[lili-li-cpu]

并且，我后续下载了您的bert4torch最新版本，然后尝试运行casrel文件，出现了这个问题
Traceback (most recent call last):
File "D:/PycharmProjects/1221bert4torch/bert4torch-master/examples/relation_extraction/task_relation_extraction_CasRel.py", line 190, in
train_model = Model().to(device)
File "D:/PycharmProjects/1221bert4torch/bert4torch-master/examples/relation_extraction/task_relation_extraction_CasRel.py", line 141, in init
self.bert = build_transformer_model(config_path, checkpoint_path)
File "D:\anaconda3\envs\DL2\lib\site-packages\bert4torch\models.py", line 2128, in build_transformer_model
transformer = MODEL(**configs)
File "D:\anaconda3\envs\DL2\lib\site-packages\bert4torch\models.py", line 423, in init
self.embeddings = BertEmbeddings(**self.get_kw('vocab_size', 'embedding_size', 'hidden_size', 'max_position', 'segment_vocab_size',
File "D:\anaconda3\envs\DL2\lib\site-packages\bert4torch\layers.py", line 457, in init
self.word_embeddings = nn.Embedding(vocab_size, embedding_size, padding_idx=pad_token_id, device=kwargs['skip_init'])
TypeError: init() got an unexpected keyword argument 'device'
我此时用的版本仍是bert4torch==0.2.8和torch4keras==0.0.7.post3
这个是什么原因呀

[Tongjilibo]

callback问题是因为新版本新建了callback.py，原有的snipets中不再支持import callback了，我刚试了下，我提供的示例是可以跑的，还有建议用最新版~