Training exits before the while loop criteria is met

[DeLeonOscar]

Hello Idree,

I am running a new training and the code is exiting before meeting the criteria, why is this happening and how can I have a complete training without being stopped too frequently?

Thank you.

[Michael-Fuu]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

[Michael-Fuu]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

[Oliverbihop]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

[Michael-Fuu]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

[WangJuan6]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

[Michael-Fuu]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

simply use vae.py to retrain vae network, don't use the model parameters under autoencoder/model/current

[WangJuan6]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

simply use vae.py to retrain vae network, don't use the model parameters under autoencoder/model/current

Thank you very much!

[Akanksham12]

can you plz tell what system configuration is needed to run this?
my system configuration is quite low :(
Intel(R) Core(TM) i3-7020U CPU @ 2.30GHz 2.30 GHz
8.00 GB
64-bit operating system, x64-based processor

@WangJuan6 @Oliverbihop @Michael-Fuu

[july000]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

simply use vae.py to retrain vae network, don't use the model parameters under autoencoder/model/current

After using the vae.py to retrain vae network, the sigma in the VariationalEncoder class is calculated inf.
How can I solve the problem?

[MaxCrazy1101]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

simply use vae.py to retrain vae network, don't use the model parameters under autoencoder/model/current

After using the vae.py to retrain vae network, the sigma in the VariationalEncoder class is calculated inf. How can I solve the problem?

same problem, have you solved the problem yet?

[yzpppp1275]

为什么跑20多轮后就Episode: 28 , Timestep: 31494 , Reward: 167.35 , Average Reward: 210.74

Exit
WARNING: sensor object went out of the scope but the sensor is still alive in the simulation: Actor 1028 (sensor.other.collision)
WARNING: sensor object went out of the scope but the sensor is still alive in the simulation: Actor 1027 (sensor.camera.rgb)
WARNING: sensor object went out of the scope but the sensor is still alive in the simulation: Actor 1026 (sensor.camera.semantic_segmentation)

[rajesh-b4]

I also have similar problem, the training process being stopped too frequently, while I'm using carla of version 0.9.14.

a simple solution is to make x_driver.py && environment.py sleep longer time, while decrease the efficiency.

Can you explain details? Thank you

Sorry for seeing this message so late. I have solved this problem, just ignore messages I mentioned before, the true solution is retraining the vae network, as this network might output nan value when using original network parameters.

Hello, I have the same question, can you explain in detail? Thank you!

simply use vae.py to retrain vae network, don't use the model parameters under autoencoder/model/current

After using the vae.py to retrain vae network, the sigma in the VariationalEncoder class is calculated inf. How can I solve the problem?

will you able to solve this problem?