error while running "packnet"

[HyunghoNa]

Hi everyone.

when defining eval_algo at packnet.py in the following code, I confront the NoneType error

def get_eval_algo(self, task_id):
# TODO: find a better way to do this
# save and load a new model and set all params where mask > current_task + 1 to 0
torch_save_model(
self.policy,
os.path.join(self.experiment_dir, "tmp_model.pth"),
cfg=self.cfg,
)
eval_algo = safe_device(
eval(self.cfg.lifelong.algo)(
eval(self.cfg.benchmark_name)().n_tasks, self.cfg
),
self.cfg.device,
)
model_state_dict, _, _ = torch_load_model(
os.path.join(self.experiment_dir, "tmp_model.pth")
)
eval_algo.policy.load_state_dict(model_state_dict)

Does anyone resolve this issue?