Can't reproduce the results

[dokhyi]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you.

Best

[12kobe]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you.

Best

Hi, I have a question for you. I get this error when I try to run this program（obstacle_avoidance_carla.py）. Do you know how to solve this problem?

[YYS188]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you.

Best

Can I get your qq or wx to ask you some questions when I reproduce the paper? Thank you, from Beijing Institude Of Technology.

[12kobe]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you. Best

Can I get your qq or wx to ask you some questions when I reproduce the paper? Thank you, from Beijing Institude Of Technology.

Hello, do you know where the paper corresponding to this code can be downloaded? Or do you mind telling me what the title of this paper is?Thank you!

[YYS188]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you. Best

Can I get your qq or wx to ask you some questions when I reproduce the paper? Thank you, from Beijing Institude Of Technology.

Hello, do you know where the paper corresponding to this code can be downloaded? Or do you mind telling me what the title of this paper is?Thank you!

Are you Chinese?If you are,we can commuicate by other tools.

[dokhyi]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you. Best

Can I get your qq or wx to ask you some questions when I reproduce the paper? Thank you, from Beijing Institude Of Technology.

Hello, do you know where the paper corresponding to this code can be downloaded? Or do you mind telling me what the title of this paper is?Thank you!

Sorry for not getting back to you sooner，you can check this paper here https://doi.org/10.1016/j.trc.2021.103452

[Cisaqiu]

你好，我在跑这个项目时提示错误，并且确实在env_v1_config没有看到报错的变量，因为我本身不是很懂代码，提错问题请见谅

[Arch-suzuki-MB]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you. Best

Can I get your qq or wx to ask you some questions when I reproduce the paper? Thank you, from Beijing Institude Of Technology.

I also try to reproduce the result like it provided in the pictures/ DQN_with_fixed _target.gif) but I didn't achived does he provided to you i model that shown this result if so can you please share it whith me

[Arch-suzuki-MB]

@YoungYoung619 I found your article very interseting and I coded but it' not giving me the same resullt as it's mentionned in the article so I execut your codes but it's not providing the same result the agent aren't able to drive and control the car . So please, can you provide your models DQN ,DDQN ,Dueling DQN , Proritized replay DQN that it's used to control the car like in the gif pictures such as (pictures/ DQN_with_fixed _target.gif) I need it to my article and thanks a lot @dokhyi, @YYS188 if it's possible to answer and thanks a lot for anyone whom trying to help in this issues.

[dokhyi]

你好，我在跑这个项目时提示错误，并且确实在env_v1_config没有看到报错的变量，因为我本身不是很懂代码，提错问题请见谅

It appears that the synchronization mode in Carla is not properly configured. It seems like the synchronization mode in Carla 0.9.5 may not serve much purpose, as I vaguely recall attempting to turn it off at some point. However, as it has been quite some time, I cannot be entirely certain.

[dokhyi]

@YoungYoung619 I found your article very interseting and I coded but it' not giving me the same resullt as it's mentionned in the article so I execut your codes but it's not providing the same result the agent aren't able to drive and control the car . So please, can you provide your models DQN ,DDQN ,Dueling DQN , Proritized replay DQN that it's used to control the car like in the gif pictures such as (pictures/ DQN_with_fixed _target.gif) and thanks a lot @dokhyi, @YYS188 if it's possible to answer and thanks a lot for anyone whom trying to help in this issues.

At the time, I attempted to train PRDQN and was able to train an agent that approached collision-free operation in an environment with stationary obstacles. The training process converged quite quickly. However, when attempting to use Dueling DQN, I found that it did not converge at all. Additionally, I couldn't find any code relevant to risk assessment as mentioned in the paper within this code repository. It's possible that I may have missed it, so please let me know if it exists.

[wolf212121j]

Hi there, I found your work very interesting and I appreciate the efforts in open-sourcing the codes. However, I have a hard time reproducing the results based on the unmodified code you provided. I tried several times in the simpler scenario 1, and even over 2000 iterations, the agent ( DQN_With_Fixed_Q_Targets ) did not achieve the results shown in your paper or code (pictures/DQN_with_fixed_target.gif). Could you kindly provide the model that are able to reproduce the results in the gifs you provided (pictures/ DQN_with_fixed _target.gif)? Thank you.

Best

Hello, do you know where I can find the training code? I found that it is not written in the code