crowdAI is shutting down - please read our blog post for more information
Reinforcement learning environments with musculoskeletal models
By Stanford Neuromuscular Biomechanics Laboratory
about 2 years ago
The simulator often gets stuck when calling reset() function.
I traced the function calls starting from reset() and it stops at
xs = np.random.uniform(1.0, 5.0, num_obstacles) # line 235 in generate_env() of run.py
and does not return.
Currently, I am working on Python 3.6.2 and Numpy 1.13.1.
Does anyone have similar problems?
about 2 years ago |
I see the problem when I create multiple environments using multi-processes.
For now, my best guess is that np.random.uniform() is not thread-safe.