I observe the missing part and unexpected part,and find that the missing part key has a “encoder” prefix, and has an extra key named “encoder.encoder.embed_tokens.weight”.besides these,they are almost the same. for exmaple,if there is an unexpected key”encoder.block.0.layer.0.SelfAttention.k.weight”,then there is a missing key “encoder.encoder.block.0.layer.0.SelfAttention.k.weight”.

so I add the prefix before I load the weight into static_dict.and the final script is as follows:

   # Check if the input .safetensors file exists
    if not os.path.exists(safetensors_path):
        print(f'Error: The .safetensors file does not exist: {safetensors_path}')
        sys.exit(1)

    # Check if the directory for the output .ckpt file exists
    ckpt_dir = os.path.dirname(ckpt_path)
    if not os.path.exists(ckpt_dir):
        os.makedirs(ckpt_dir)

    print(f'Loading weights from: {safetensors_path}')

    try:
        # Load the .safetensors file
        weights = load_file(safetensors_path, device='cpu')

        # # Ensure weights are in state_dict format
        # if isinstance(weights, dict) and 'state_dict' in weights:
        #     state_dict = weights['state_dict']
        # else:
        #     # If weights itself is a state_dict
        nw={}
        for key,value in weights:
            nk="encoder."+key
            nw[nk]=value


        # Add prefix to state_dict keys

        # Prepare the data to be saved in .ckpt format
        # Adding state_dict under 'state_dicti' key to match typical .ckpt structure
        checkpoint_data = {'state_dict': nw}

        # Save the state_dict to .ckpt file
        torch.save(checkpoint_data, ckpt_path)
        print(f'Weights successfully saved to: {ckpt_path}')

but as I mentioned,there is a missing key,so still cant work
but the new error message says that I should pass a “strict=false” to the load_static_dict function.
but HOW?
I found this in pytorch_lightning/core/module.py:

    @property
def strict_loading(self) -> bool:
    """Determines how Lightning loads this model using `.load_state_dict(..., strict=model.strict_loading)`."""
    # We use None as the default internally to determine whether the user has set a value
    return self._strict_loading in (None, True)

so I add just one line in the initialize of the model:

self._strict_loading = false

and it works!

it is really hard for me to figure out all this, and I hope it may help those who has the same problem as me.

I also see solutiontext like:

self.model.load_state_dict(dict([(n, p) for n, p in checkpoint['model'].items()]), strict=False)

but since the original code has defined functions to load checkpoint, I don’t know how to add this into it.
but add one parameter to the model is easy.