-
Notifications
You must be signed in to change notification settings - Fork 20
Open
Description
Hi, I was getting the following error when running the evaluation script bash scripts/eval_vllm_ray.sh 0,1. I made no change to the source code and I directly loaded the untrained openVLA checkpoint. The GPU I used is NVIDIA H200. Seems that this error is from multithread processing.
Traceback (most recent call last):
File "/gpfs/self-evolve/vlarl/run_libero_eval_vllm.py", line 373, in <module>
eval_libero()
File "/home/.conda/envs/vlarl/lib/python3.10/site-packages/draccus/argparsing.py", line 228, in wrapper_inner
response = fn(cfg, *args, **kwargs)
File "/gpfs/self-evolve/vlarl/run_libero_eval_vllm.py", line 321, in eval_libero
next_obs, rewards, dones, infos = envs.step(actions)
File "/gpfs/self-evolve/vlarl/ppo/envs/libero_env.py", line 146, in step
return self._step_impl(action, **kwargs)
File "/gpfs/self-evolve/vlarl/ppo/envs/libero_env.py", line 304, in _step_impl
obs_np_list, reward_np_list, done_np_list, info = self.env.step(normalized_action.tolist(), id=valid_task_ids)
File "/gpfs/self-evolve/vlarl/ppo/envs/venv.py", line 803, in step
env_return = self.workers[j].recv()
File "/gpfs/self-evolve/vlarl/ppo/envs/venv.py", line 446, in recv
result = self.parent_remote.recv()
File "/home/.conda/envs/vlarl/lib/python3.10/multiprocessing/connection.py", line 250, in recv
buf = self._recv_bytes()
File "/home/.conda/envs/vlarl/lib/python3.10/multiprocessing/connection.py", line 414, in _recv_bytes
buf = self._recv(4)
File "/home/.conda/envs/vlarl/lib/python3.10/multiprocessing/connection.py", line 383, in _recv
raise EOFError
EOFErrorReactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels