TypeError: _build_causal_attention_mask() missing 1 required positional argument: 'dtype'

[bigcash]

when i run command: python export_df_onnx.py
got error:
Traceback (most recent call last):
File "export_df_onnx.py", line 136, in
width=512,
File "export_df_onnx.py", line 115, in convert_to_onnx
text_encoder, check_inputs[0], check_inputs=[check_inputs[1]], strict=False
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/jit/_trace.py", line 750, in trace
_module_class,
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/jit/_trace.py", line 965, in trace_module
argument_names,
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1090, in _slow_forward
result = self.forward(*input, **kwargs)
File "/home/lingbao/.local/lib/python3.7/site-packages/transformers/models/clip/modeling_clip.py", line 728, in forward
return_dict=return_dict,
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/home/lingbao/.local/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1090, in _slow_forward
result = self.forward(*input, **kwargs)
File "/home/lingbao/.local/lib/python3.7/site-packages/transformers/models/clip/modeling_clip.py", line 637, in forward
causal_attention_mask = self._build_causal_attention_mask(bsz, seq_len).to(hidden_states.device)
TypeError: _build_causal_attention_mask() missing 1 required positional argument: 'dtype'

if i add this argument by this(line 115):
text_encoder, check_inputs[0], check_inputs=[check_inputs[1]],dtype=torch.int32, strict=False
then another error come "TypeError: trace() got an unexpected keyword argument 'dtype'"

maybe conflict version of transformers, my env: transformers==4.19.2