Ask for the multiprocessing issue in the function "evaluate_one_task_success"

[FrankSinatral]

Thanks for your great work! When I try to run the main.py training script, I have the following problem:Exception has occurred: BrokenPipeError
[Errno 32] Broken pipe
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/libero/envs/venv.py", line 425, in send
self.parent_remote.send(["reset", kwargs])
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/libero/envs/venv.py", line 708, in reset
self.workers[i].send(None, **kwargs)
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/lifelong/metric.py", line 110, in evaluate_one_task_success
env.reset()
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/lifelong/metric.py", line 191, in evaluate_multitask_training_success
success_rate = evaluate_one_task_success(cfg, algo, task_i, task_emb, i)
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/lifelong/algos/skill.py", line 310, in learn_multi_task
success_rates = evaluate_multitask_training_success(
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/lifelong/main.py", line 303, in main
s_fwd, l_fwd, kl_loss, ce_loss, embedding_loss = meta_policy.learn_multi_task(meta_dataset, benchmark, result_summary, cfg.use_wandb)
File "/mnt/datadisk1/fanke/Projects/Lotus/lotus/lifelong/main.py", line 464, in
main()
BrokenPipeError: [Errno 32] Broken pipe，I guess there are some problems inside the class SubprocVectorEnv， how should I fix the problem, Thanks!