Size mismatch while loading the pretrained weights for retraining the layoutlmv2

[deepanshudashora]

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
[<ipython-input-23-2ca38188b5f5>](https://localhost:8080/#) in <module>()
      4 
      5 model = LayoutLMv2ForTokenClassification.from_pretrained('/content/drive/MyDrive/resume_experiment/checkpoints',
----> 6                                                                       num_labels=len(labels))
      7 
      8 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

1 frames
[/usr/local/lib/python3.7/dist-packages/transformers/modeling_utils.py](https://localhost:8080/#) in from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)
   1363                 pretrained_model_name_or_path,
   1364                 ignore_mismatched_sizes=ignore_mismatched_sizes,
-> 1365                 _fast_init=_fast_init,
   1366             )
   1367 

[/usr/local/lib/python3.7/dist-packages/transformers/modeling_utils.py](https://localhost:8080/#) in _load_state_dict_into_model(cls, model, state_dict, pretrained_model_name_or_path, ignore_mismatched_sizes, _fast_init)
   1510         if len(error_msgs) > 0:
   1511             error_msg = "\n\t".join(error_msgs)
-> 1512             raise RuntimeError(f"Error(s) in loading state_dict for {model.__class__.__name__}:\n\t{error_msg}")
   1513 
   1514         if len(unexpected_keys) > 0:

RuntimeError: Error(s) in loading state_dict for LayoutLMv2ForTokenClassification:
	size mismatch for classifier.weight: copying a param with shape torch.Size([47, 768]) from checkpoint, the shape in current model is torch.Size([48, 768]).
	size mismatch for classifier.bias: copying a param with shape torch.Size([47]) from checkpoint, the shape in current model is torch.Size([48]).

This is the issue I am facing when I am loading the previously trained weights of my model to retrain it on new data, you can see in the code how I changed the loading of pretrained weights, is it the correct way of doing that?

from transformers import LayoutLMv2ForTokenClassification, AdamW
import torch
from tqdm.notebook import tqdm

**model = LayoutLMv2ForTokenClassification.from_pretrained('/content/drive/MyDrive/resume_experiment/checkpoints',
                                                                      num_labels=len(labels))**

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model.to(device)
optimizer = AdamW(model.parameters(), lr=5e-5)

global_step = 0
num_train_epochs = 10

#put the model in training mode
model.train() 
for epoch in range(num_train_epochs):  
   print("Epoch:", epoch)
   for batch in tqdm(train_dataloader):
        # get the inputs;
        input_ids = batch['input_ids'].to(device)
        bbox = batch['bbox'].to(device)
        image = batch['image'].to(device)
        attention_mask = batch['attention_mask'].to(device)
        token_type_ids = batch['token_type_ids'].to(device)
        labels = batch['labels'].to(device)

        # zero the parameter gradients
        optimizer.zero_grad()
        
        # forward + backward + optimize
        outputs = model(input_ids=input_ids,
                        bbox=bbox,
                        image=image,
                        attention_mask=attention_mask,
                        token_type_ids=token_type_ids,
                        labels=labels) 
        loss = outputs.loss
        
        # print loss every 100 steps
        if global_step % 100 == 0:
          print(f"Loss after {global_step} steps: {loss.item()}")

        loss.backward()
        optimizer.step()
        global_step += 1

model.save_pretrained("/content/drive/MyDrive/resume_experiment/checkpoints_2")