TypeError: expected str, bytes or os.PathLike object, not dict

[llingss]

Hello author, I am currently running this code and I have a small question to ask you,
I created the corresponding virtual environment according to the first method in the readme document and successfully installed deepinterpolation. When running the command "python exemplar_tiny-epys_training. py", an error message was displayed: "TypeError: expected str, bytes or os.PathLike object, not dict"， I tried to convert generator-paramam into a JSON file in the exed_tiny_ephys_training.py file,
The same error occurred with generator_test_maram later on, and I also used the same handling method
"# Write generator_test_maram parameter to JSON file
with open('generator_test_param.json', 'w') as f:
json.dump(generator_test_param, f)
#Pass JSON file path instead of dictionary
test_generator = EphysGenerator('generator_test_param.json')”，"
But the same processing method as the previous two steps still cannot solve the problem of "unet_single_cephys_1024 ({}), traininglparam". May I ask if there is a problem with my processing method? What caused this error?