Error while setting ray/Impala configuration

🎈 Using Streamlit
1

Hello there, I’m getting error trying to init my trainer model with ray.
The error is " __deepcopy__() is not a valid Streamlit command.", don’t know why streamlit is trying to deepcopy when streamlit is never called in that part of the code

image
image1179×727 38.2 KB

Here is the code:

   ray.init(local_mode=True, include_dashboard=False, ignore_reinit_error=True)
   config = impala.impala.DEFAULT_CONFIG.copy()

    config["env"] = SimulationEnvironment
    if env_config:
        config["env_config"] = env_config
    config["num_workers"] = 1
    config["sample_async"] = False
    config["num_gpus"] = 0
    config["rollout_fragment_length"] = 48
    config["train_batch_size"] = config["rollout_fragment_length"] * 5
    trainer = impala.impala.ImpalaTrainer(config)      <--- Error
    trainer.restore(
        "../reinforcement_learning/models/IMPALA/IMPALA_SimulationEnvironment_0_2020-10-08_20-55-26efyv3ax6/checkpoint_2200/checkpoint-2200")

Any ideas?

Thanks!