Comments (8)
The things that I should do next:
-
Optimize
broadcasting
until it beats numpy/pytorch
The current implementation of broadcasting depends on BLAS/CUDA Operation, failing to parallelize.
I guess I need to look for other ways of implementation.... -
Support Half(FP16)
i gonna call CFFI by myself or something -
Save and Restore Models.
support compatibility with npz
from cl-waffe.
The first version won't be released until the above is completed if not all.
from cl-waffe.
I'm working on inlining call-forward/call-backward.
This works really nicely.
from cl-waffe.
I've inlined call-forward/call-backward, #120
from cl-waffe.
とりあえずやること:
ドキュメントの編集 (Important)
kernel.lispの最適化 (it's done?)
npzのパーサーをかく (turned to be unnecessary)
from cl-waffe.
Anyway, the rest things should be done until the first version release are the following:
- Prepare more documents and tutorials for those who are new to cl-waffe
- More impls of NN: (e.g. RNN/LSTM/GRU (make them 100x faster!), CNN/AvgPooling MaxPooling/, Attentions, faster embedding, Normalizations, More various activations etc...)
- To achieve 2., the speed of iterations is important.
from cl-waffe.
- float16
- !view, which works like torch's one.
- refactoring whole codes
from cl-waffe.
build documentations on mkdocs
from cl-waffe.
Related Issues (19)
- The backpropagation of !split is undefined. HOT 2
- Split large files into small HOT 1
- !argmax/!argmin requires tensor to be copied.
- To add: Dispatch nodes based on a given tensor’s backend
- The performance of call-forward and call-backward. HOT 4
- call-forward, call-backward is reducible HOT 4
- Support GPU
- (setf !aref) is ugly
- Add LSTM/GRU HOT 1
- Add Conv/Pooling
- Save and restore the weights of model HOT 1
- Add: DataTypeまわりの機能がもっと欲しい HOT 1
- Implement !expands
- Sampling beta distributions is failed with alpha=1.0, beta=1.0
- Add: (with-config ~) macro
- 重みの初期化手法について HOT 1
- Gradient explosion of Seq2Seq (possibly related to linearlayer/embedding) HOT 4
- Allowing other options to use matrices in cl-waffe HOT 1
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 cl-waffe.