---
tags:
- Taxi-v3
- q-learning
- reinforcement-learning
- custom-implementation
- deep-rl-course
model-index:
- name: q-Taxi-v3
  results:
  - task:
      type: reinforcement-learning
      name: reinforcement-learning
    dataset:
      name: Taxi-v3
      type: Taxi-v3
    metrics:
    - type: mean_reward
      value: 7.52 +/- 2.73
      name: mean_reward
      verified: false
---
# Q-Learning Agent playing Taxi-v3

This is a trained Q-Learning agent for
Hugging Face Deep Reinforcement Learning Course Unit 2.

## Evaluation

- Environment: Taxi-v3
- Mean reward: 7.52
- Standard deviation: 2.73
- Result (`mean_reward - std_reward`): **4.79**
- Required course result: **4.5**
- Evaluation episodes: 100

## Compatibility note

The Q-table was trained and evaluated with `Taxi-v4` in the current
Gymnasium release because `Taxi-v3` is deprecated there.

The Hugging Face Deep RL Course identifier remains `Taxi-v3`.