采用args或kwargs形式输入时，如果args或kwargs输入的变量是list或dict，会报错

[zgcr]

典型示例如下：
from calflops import calculate_flops
flops, macs, params = calculate_flops(model=net,
kwargs={
'x':[C3,C4,C5],
},
output_as_string=True,
output_precision=3,
print_results=False,
print_detailed=False)
print(f'1111, flops: {flops}, macs: {macs}, params: {params}')
若C3，C4，C5都为torch tensor，运行后会出现以下错误：
Traceback (most recent call last):
File "/root/code/SimpleAICV_pytorch_training_examples/SimpleAICV/detection/models/fpn.py", line 107, in
flops, macs, params = calculate_flops(model=net,
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/miniconda3/lib/python3.12/site-packages/calflops/flops_counter.py", line 154, in calculate_flops
kwargs[key] = value.to(device)
^^^^^^^^
AttributeError: 'list' object has no attribute 'to'

类似地，若
kwargs={
'batch_prompts':input_prompts,
},
假如
input_prompts = {
'prompt_point': input_prompt_points,
'prompt_box': input_prompt_boxs,
'prompt_mask': input_prompt_masks,
}
是一个dict，dict中每个key对应的value是一个tensor，也会报相同错误。