jkulhanek / visual-navigation-agent-pytorch

Target-driven Visual Navigation in Indoor Scenes using Deep Reinforcement Learning implemented in PyTorch
MIT License
63 stars 18 forks source link

Error after training starts: "type object got multiple values for keyword argument 'max_t' " #4

Open TomasMendozaHN opened 4 years ago

TomasMendozaHN commented 4 years ago

I am using Ubuntu16.04. I follow the steps you specified to start training. Shortly after the text "training started" shows, i get the error "type object got multiple values for keyword argument 'max_t'". You can see my terminal output below:

Status: Downloaded newer image for kulhanek/visual-navigation-agent-pytorch:latest Training started Traceback (most recent call last): File "/app/train.py", line 39, in t.run() File "/app/agent/training.py", line 243, in run self.threads = [_createThread(i, task) for i, task in enumerate(branches)] File "/app/agent/training.py", line 243, in self.threads = [_createThread(i, task) for i, task in enumerate(branches)] File "/app/agent/training.py", line 241, in _createThread **self.config) TypeError: type object got multiple values for keyword argument 'max_t'

What can i do? I didn't change anything in your code, just cloned it from github and executed it like your instructions.

pushkalkatara commented 4 years ago

Hi @tmendozarias , were you able to resolve this issue?

cenotence commented 4 years ago

I think this problem is occurred because max_t assigned twice in training.py file.

So comment out or deleat Ln 29 'max_t' :5, in training.py, where assign default_config.

DEFAULT_CONFIG = { 'saving_period': 10 ** 6 // 5, 'checkpoint_path': 'model/checkpoint-{checkpoint}.pth', 'grad_norm': 40.0, 'gamma': 0.99, 'entropy_beta': 0.01,

'max_t': 5,

}

this worked for me.