Spaces:

---

huggingface-projects

/

Deep-Reinforcement-Learning-Leaderboard

like 383

Running on CPU Upgrade

I am also having the same issue with frozenLake-v1-4x4-noSlippery => https://huggingface.co/mahir05/q-FrozenLake-v1-4x4-noSlippery

I'm facing the same issue after all the problems I had using the LunarLander-v2 and finally used LunarLander-v3 but using that name.

Guys, I just noticed that in Unit 0, it was mentioned that the leaderboard is no longer operational. But if you're looking for a way to check if your agent has been evaluated for getting a certificate, you can just check the link below and see if you've passed the assignment.
https://huggingface.co/spaces/ThomasSimonini/Check-my-progress-Deep-RL-Course

Oh. Thanks a lot. I somehow missed that.

I uploaded my model but can not find my assignment progress in this link https://huggingface.co/spaces/ThomasSimonini/Check-my-progress-Deep-RL-Course . What should I do?

I have implemented the Unit 1 model (ppo-LunarLander-v3) and got a result of >= 200. But After I pushed to the Hugging Face Hub, I can't find the progress as passed for Unit 1 yet.

@ishadyaAP I think the issue is with your metadata (README.md).

library_name: stable-baselines3
tags:
  - LunarLander-v3
  - deep-reinforcement-learning
  - reinforcement-learning
  - stable-baselines3
model-index:
  - name: PPO
    results:
      - task:
          type: reinforcement-learning
          name: reinforcement-learning
        dataset:
          name: LunarLander-v3
          type: LunarLander-v3
        metrics:
          - type: mean_reward
            value: 253.65 +/- 20.95
            name: mean_reward
            verified: false

Here, it should be LunarLander-v2 in place of LunarLander-v3. Though I am not certain that it'll work, but you can give it a try. Or you can just re-train using the LunarLander-v2 environment.

@mahir05 Many Thanks to you. Changing the README.md worked. I changed all 'LunarLander-v3' terms to 'LunarLander-v2'.

I was also thinking that is the issue, but couldn't find where the error is. I think my colab version doesn't let me to train the model using LunarLander-v2, it gives an error. So I trained it using LunarLander-v3.

This will be helpful to others who train the model using LunarLander-v3 instead of LunarLander-v2. After pushing the model to the HuggingFace hub go to your code and open the README.md file and do the changes I mentioned.

Thanks to @mahir05 again

Thanks @mahir05 ! 🙏

Your solution worked perfectly! I trained my model using LunarLander-v3 (because v2 was deprecated in the latest gymnasium) and couldn't find it on the leaderboard or progress checker.

After changing all LunarLander-v3 to LunarLander-v2 in my README.md metadata, it finally got recognized!

This should definitely be added to the course documentation since many users will face this issue with newer gymnasium versions.

Thanks again for the helpful tip! 🎉