You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
if you want to harness the power of multiprocessing we have to debug what happens if self.multiprocess is set to true.
Some potential fixes include:
1. could this have anything to do with the device at all????
2. few things to try: 1. us the sub proc vec env with one if self.num_parralel_envs = 1,
3. simple script to test out a simple env, and test out subprocvecenv class
4. simple script to test out a simple env, and test out subprocvecenv class
5. go to github to stable baselines => got to stable baselines and search in their issues you might.
6. write a script to reproduce the error create a simple script that recreates the error.
def create_new_image_envs(self): if self.multiprocess: self.envs = VecPyTorch( SubprocVecEnv([self.make_env(self.seed + i) for i in range(self.num_parallel_envs)], 'fork'), self.device) else: self.envs = VecPyTorch(DummyVecEnv([self.make_env(self.seed + i) for i in range(self.num_parallel_envs)]), self.device)
The text was updated successfully, but these errors were encountered:
if you want to harness the power of multiprocessing we have to debug what happens if self.multiprocess is set to true.
Some potential fixes include:
1. could this have anything to do with the device at all????
2. few things to try: 1. us the sub proc vec env with one if self.num_parralel_envs = 1,
3. simple script to test out a simple env, and test out subprocvecenv class
4. simple script to test out a simple env, and test out subprocvecenv class
5. go to github to stable baselines => got to stable baselines and search in their issues you might.
6. write a script to reproduce the error create a simple script that recreates the error.
def create_new_image_envs(self): if self.multiprocess: self.envs = VecPyTorch( SubprocVecEnv([self.make_env(self.seed + i) for i in range(self.num_parallel_envs)], 'fork'), self.device) else: self.envs = VecPyTorch(DummyVecEnv([self.make_env(self.seed + i) for i in range(self.num_parallel_envs)]), self.device)
The text was updated successfully, but these errors were encountered: