RL error in nb in Section 17, Unit 1

The cell 11, where we have: 

#Call the run function and pass the dataframe and hyperparameters
run(bars5m, rl_config)
 

it starts to execute, and then throws the error few times. It's a cryptic error.
2023-02-12 04:34:08.018131: W tensorflow/core/kernels/data/model_dataset_op.cc:205] Optimization loop failed: Cancelled: Operation was cancelled

Why? 

Now, i got the message that its limited to 15 episodes, although the  error was thrown during the first and on each 2-3 subsequent eposides. This is an expensive course, and I believe you should had been bit more generous with the eposides. At least 50 times to that some conclusion could be drown. Also, even if it stops with a limit of 15 eposodes, why should it throw error? I believe the error is something else. Any way, I feel , here the project is a bit fudged and unporifessonally handeled. Please advise and help.

Thanks!

 Hello Antoni, 

  1. This warning is because of the absence of a GPU. This is nothing to be worried about. If GPU is available, the optimisation will kick in through parallel computing.



    I ran the code on a local system with GPU and I can successfully run it without any warnings. 

    This warning can be suppressed by adding the following piece of code before importing TensorFlow. 
# To suppress GPU-related warnings
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'

2) The TEST_MODE was set to 'True' in rl_config due to resource constraints. The training stops after 15 episodes in test mode. The training wouldn't stop if TEST_MODE is false. We recommend you try this in your local system. You can use the notebook in the zip file. Here is the link to download the zip file. 

I hope this helps!