IntelLabs / coach

Reinforcement Learning Coach by Intel AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms
https://intellabs.github.io/coach/
Apache License 2.0
2.32k stars 459 forks source link

Install coach failed #271

Closed fangchuan closed 5 years ago

fangchuan commented 5 years ago

Hi, recently I had the same errors when I try to install the coach, whether I used the pip3 install -e . or pip3 install rl_coach , these command outputs same error information:

`(intel_coach) fc@fc-msi:~/toolkits/ReinforcementLearning/Coach/intel_coach/coach$ pip3 install -e . Obtaining file:///home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/coach Complete output from command python setup.py egg_info: Traceback (most recent call last): File "", line 1, in File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/setuptools/init.py", line 18, in import setuptools.version File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/setuptools/version.py", line 1, in import pkg_resources File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2927, in @_call_aside File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2913, in _call_aside f(*args, **kwargs) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2952, in _initialize_master_working_set add_activation_listener(lambda dist: dist.activate()) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 956, in subscribe callback(dist) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2952, in add_activation_listener(lambda dist: dist.activate()) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2515, in activate declare_namespace(pkg) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2097, in declare_namespace _handle_ns(packageName, path_item) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2047, in _handle_ns _rebuild_mod_path(path, packageName, module) File "/home/fc/toolkits/ReinforcementLearning/Coach/intel_coach/lib/python3.5/site-packages/pkg_resources/init.py", line 2066, in _rebuild_mod_path orig_path.sort(key=position_in_sys_path) AttributeError: '_NamespacePath' object has no attribute 'sort'

` Is there anybody who can help me? I'm hurry, thanks

gal-leibovich commented 5 years ago

Hi, there's seems to be something messy with your environment. Have you tried installing in a clean virtual_env?

fangchuan commented 5 years ago

Hi, thanks @galleibo-intel , I tried to install coach in a clean virtual_env and it works. However, when I want to run the preset of CARLA_DDPG, it always crashes when it comes to "Starting to improve simple_rl_graph task index 0": ` coach -r -p CARLA_DDPG Please enter an experiment name: CARLA_DDPG

Creating graph - name: BasicRLGraphManager Creating agent - name: agent simple_rl_graph: Starting heatup Heatup - Name: main_level/agent Worker: 0 Episode: 1 Total reward: -5929.35 Steps: 999 Training iteration: 0 Heatup - Name: main_level/agent Worker: 0 Episode: 2 Total reward: -607.95 Steps: 1097 Training iteration: 0 Starting to improve simple_rl_graph task index 0 Segment fault.... ` I have seen the worker_0.simple_rl_graph_0.json , is it possible caused by the memory size of replay buffer? The original size is 1M, while I only have 8G memory on my computer.

gal-leibovich commented 5 years ago

Generally, it might pose a problem, but not so early in the process, as the replay buffer is not filled yet. Closing this issue, as the issue was resolved.