masterkey2000

> If you don't pass the observation_action_splitter to DQN then the observation contains both the observation and the mask. And the mask is bool which cannot be sampled. Not sure...

> Also can you make sure you have the latest version since sample_nest_spec can now handle tf.bool https://github.com/tensorflow/agents/blob/master/tf_agents/specs/tensor_spec.py#L380 the tf-agents version I use is 0.12.0

Hello, I followed the instructions from @Bahador-Bakhshi but I get an error when running ``` agent = dqn_agent.DqnAgent( train_env.time_step_spec(), train_env.action_spec(), q_network=q_net, optimizer=optimizer, observation_and_action_constraint_splitter=environment.observation_action_splitter, td_errors_loss_fn=common.element_wise_squared_loss, #train_step_counter=train_step_counter) train_step_counter=global_step) ``` The Error is:...

@sguada: What do you mean? I did it exactly as @Bahador-Bakhshi suggested, or did I miss something?