FastAI.jl
AlphaZero.jl
FastAI.jl | AlphaZero.jl | |
---|---|---|
3 | 2 | |
584 | 1,217 | |
0.5% | - | |
4.0 | 3.5 | |
2 months ago | about 2 months ago | |
Julia | Julia | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
FastAI.jl
-
Python vs Julia
You should definitely go with Julia. It has steeper learning curve than python, but it is way more powerful. As for the ecosystem, you shouldn't worry about that much: DataFrames.jl and friends is way better than pandas, MLJ.jl (https://github.com/alan-turing-institute/MLJ.jl) and FastAI.jl(https://github.com/FluxML/FastAI.jl) are great frameworks for regular ML and deepnet. And if at any point you get a feeling that you need some python library, you can always plug it in with PyCall.jl(https://github.com/JuliaPy/PyCall.jl).
-
Flux vs. TensorFlow
Flux is nice, the API is very simple but it is lacking several utilities that PyTorch or TensorFlow have. Fortunately, [FastAI](https://github.com/FluxML/FastAI.jl) is adding these parts as a additional package. I hope very soon I could recommend the Deep Learning ecosystem in Julia without any doubt.
-
Is Julia Ready for Deep Learning
There’s apparently an ongoing effort to port fast.ai v2 to Julia at https://github.com/FluxML/FastAI.jl - there some virtual standups on YouTube tracking progress, see https://youtu.be/3g0vjDA0PBU for example.
AlphaZero.jl
-
Show HN: ChessCoach – A neural chess engine that comments on each player's moves
Could using something like AlphaZero.jl make it more efficient?
https://github.com/jonathan-laurent/AlphaZero.jl
-
Ask HN: What are some tools / libraries you built yourself?
As a researcher in machine learning, I wanted to explore applications of Deepmind’s AlphaZero algorithm beyond board games (such as in automated theorem proving or chemical synthesis).
However, I noticed that existing open-source implementations of AlphaZero mostly consisted in complex C++ codebases that are highly specialized for specific games (eg. Leela Zero and LC0). Accessible Python implementations could be found but they were usually too slow to do anything useful on limited computing power.
Seeing this, I built AlphaZero.jl: https://github.com/jonathan-laurent/AlphaZero.jl
AlphaZero.jl is written in Julia and it is consistently one to two orders of magnitude faster than competing Python alternatives, while being equally simple and flexible. I just released a new version a few days ago with many new features (support for distributed computing, support for arbitrary MDPs...).
If you are a student, a researcher or a hacker curious about AlphaZero, please consider having a look!
What are some alternatives?
MLJ.jl - A Julia machine learning framework
lowdefy - The config web stack for business apps - build internal tools, client portals, web apps, admin panels, dashboards, web sites, and CRUD apps with YAML or JSON.
TensorFlow.jl - A Julia wrapper for TensorFlow
vaku - vaku extends the vault api & cli
PyCall.jl - Package to call Python functions from the Julia language
fselect - Find files with SQL-like queries
julia - The Julia Programming Language
ChessCoach - Neural network-based chess engine capable of natural language commentary
model-zoo - Please do not feed the models
rupy - HTTP App. Server and JSON DB - Shared Parallel (Atomic) & Distributed
NaiveGAflux.jl - Evolve Flux networks from scratch!
Lila - ♞ lichess.org: the forever free, adless and open source chess server ♞ [Moved to: https://github.com/lichess-org/lila]