Comments (3)
Yes, you can use this example:
https://github.com/MushroomRL/mushroom-rl/blob/dev/examples/atari_dqn.py
you have just to change the DQN/DDQN/similar algorithms with your implementation.
You will need to change the code, particularly at lines 328-363 where the algorithms are constructed.
If you are talking about mushroom-rl-benchmarking, unfortunately, I have again to say that it's still not quite ready.
If you need a better explanation of the code, you can look at the simplified version in the tutorial:
https://mushroomrl.readthedocs.io/en/latest/source/tutorials/tutorials.3_deep.html
from mushroom-rl.
I mean more in terms of hyperparams and changing the algorithm. Is it easy to configure other algorithms for Atari?
from mushroom-rl.
The hyperparameter search toolbox is not part of mushroom at all. You can combine Mushroom structure with other optimizers (e.g. Optuna) to achieve this goal. But we don't give any support on this aspect, by the design of MushroomRL.
you can MushroomRL Benchmark to:
- run standard benchmarks to get a measure of the implementation of mushroom algorithms
- create new benchmarks and compare new algorithms with the existing ones.
It is not a tool to do RL research (if you exclude the research in the benchmarking field).
If you want to use Mushroom for your research, we suggest having a look at:
- the tutorials (have a look at the four tutorials on the left)
- the examples, maybe starting from the more simple ones such as the simple chain, the plotting and normalization and the acrobot.
- Finally, you can look at the atari linked above, which can be the starting point of your work.
Remember that the idea of Mushroom is to provide a clean, organized, and extensible platform to do RL research. By design, it is not a tool to launch easily parameter searches with a configuration file. Other libraries are better suited for this task.
With Mushroom, it is easier to develop new algorithms and reuse existing code in your novel approaches.
from mushroom-rl.
Related Issues (20)
- Can't install package HOT 4
- suspected memory leak HOT 8
- How to train an agent in one environment and use it on another slightly different envoirnment HOT 3
- dynaq agent HOT 1
- how to reproduce DQN nature paper? HOT 7
- compress frames HOT 2
- n_steps dqn performs worse. bug?
- support for new spaces HOT 2
- PPO for lunar lander [BUG] HOT 10
- Multi modal state support HOT 1
- Save and Load Agent for the Second Time HOT 2
- Tutorial for REINFORCE HOT 2
- REINFORCE with optional baseline HOT 1
- Incorrect Shape of Baseline in REINFORCE HOT 11
- QLearning Can't Train On Episodes HOT 6
- Suggestion: rename episodes_length to compute_episodes_length
- Suggestion: Add median to compute_metrics
- [solvers/dynamic_programming] Use np.linalg.solve instead of np.inv HOT 2
- [requirements.txt] Missing requirement for OpenAI gym HOT 4
- [Categorical DQN/Rainbow] Inconsistent behavior of Categorical DQN for an even number of atoms
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mushroom-rl.