search

mschrader15 / reinforcement-learning-sumo

Reinforcement Learning + traffic microsimulation (via SUMO). Uses Ray RLLIB and forces SUMO into the OpenAI Gym Framework
https://maxschrader.io/reinforcement-learning-and-sumo
45 stars 9 forks source link

requirements.txt #7

Closed TrinhTuanHung2021 closed 1 year ago

TrinhTuanHung2021 commented 2 years ago

Hello

I only run your models with python==3.8 and ray==1.11.0 Please update them in requirements.txt

After 1 day of training, my computer was out of memory with ray so I coud not have final resutls The configurations of my PC: i9, ram=32gb

TrinhTuanHung2021 commented 2 years ago

== Status == Current time: 2022-08-24 15:01:03 (running for 04:32:04.10) Memory usage on this node: 30.9/33.4 GiB: LOW MEMORY less than 10% of the memory on this node is available for use. This can cause unexpected crashes. Consider reducing the memory used by your application or reducing the Ray object store size by setting object_store_memory when calling ray.init. Using FIFO scheduling algorithm. Resources requested: 2.0/16 CPUs, 0/0 GPUs, 0.0/18.59 GiB heap, 0.0/9.3 GiB objects Result logdir: /home/osboxes/ray_results/ES-updated Number of trials: 1/1 (1 RUNNING)

TrinhTuanHung2021 commented 2 years ago

I add ray.init(object_store_memory=200**9) in rllib.py to limit the object store to use 20GB

mschrader15 commented 2 years ago

Thanks for the information @TrinhTuanHung2021 !

Do you feel comfortable making a pull request with both the requirements update as well object store update? I will merge it into the main code if so