Tkl is a machine learning (ML) platform developed in python. It is an easy-to-use lite weight ML framework. It can run on low-end pc, tablets, androide devices etc.
Tkl is a very lite weight machine learning libary it runs off the `CPU`.
*The current version (v 1.19 alplha) does not support the `GPU`
[tkl-lite]: https://drive.google.com/file/d/1EPojFMeBpxBwtM6hRvZxzTlImn7tkYaB/view?usp=sharing (Great option for android devices.)
[tkl-full]: https://drive.google.com/drive/folders/1NcWBs7T9J1kGEyddR7MuRZFWtQaFDdfl?usp=sharing (Great option for PC's)
https://trinket.io/library/trinkets/4ed462b480
from tkl import *
import random
#// setting up the inputs
input1 = NN.add_input(random.uniform(0, 1))
input2 = NN.add_input(random.uniform(0, 1))
input3 = NN.add_input(random.uniform(0, 1))
input4 = NN.add_input(random.uniform(0, 1))
#// setting up the layers
neuron1 = NN.add_neuron(input1,activations.ReLU,0.0001,alpha=2.0,gamma=-1.2)
neuron2 = NN.add_neuron(input2,activations.ReLU,0.001,alpha=2.0,gamma=-1.2)
neuron3 = NN.add_neuron(input3,activations.ReLU,0.001,alpha=2.0,gamma=-1.2)
neuron4 = NN.add_neuron(input4,activations.ReLU,0.001,alpha=2.0,gamma=-1.2)
#//seting up the outputs
output1 = NN.add_output(neuron1,[1.1,2,-2,1],1)
output2 = NN.add_output(neuron2,[0.2,2,-2,1],1)
output3 = NN.add_output(neuron3,[0.3,2,-2,1],1)| Name | Version | Description |
|---|---|---|
tkl.NN() |
1.1.4 | create simple neural networks |
tkl.activations() |
1.2.7 | activation functions for the neural network |
tkl.koala() |
1.1.3 | create deep networks |
tkl.islands() |
1.1.2 | dataset for tkl |
tkl.tkboard() |
1.1.1 | create simple graphs based on neural network |
tkl.train() |
0.1.0 | trains a Neural network/Deep network |
tkl.rhinoceros() |
0.0.1 | create neural networks & deep networks |
[1.11]:[base/root of tkl, added NN]
[1.12]:[tkl.koala was added]
[1.13]:[tkl.activations were added]
[1.14]:[tkl.koala removed for maintenance]
[1.15]:[tkl.koala was updated and added back]
[1.16]:[tkl.islands was added]
[1.17]:[tkl.islands,tkl.koala were improved]
[1.18]:[BUG fixes,fixed tkl.NN, koala improved,added tkboard]
[1.19]:[Added Train,Rhinoceros tkl.NN.add_input()/tkl.koala.add_input() array can be used]
1) download `[tkl-full].zip` from `https://drive.google.com/drive/folders/1NcWBs7T9J1kGEyddR7MuRZFWtQaFDdfl?usp=sharing`
2) unzip the file you installed (tkl-full.zip)
3) Move unzipped folder to the right path:
#ANACONDA
If you are using conda, place the unzipped files folder in the directory: `C:\ProgramData\Anaconda3\Lib`
#IDLE
If you are using IDLE, place the unzipped files folder in the directory: `C:\Users\{SYSYTEM USRNAME}\AppData\Local\Programs\Python\Python39\Lib`
1) download `[tkl-lite].py` from `https://drive.google.com/file/d/1EPojFMeBpxBwtM6hRvZxzTlImn7tkYaB/view?usp=sharing`
2) open the IDE of your choice, check if tkl is working using `import tkl`
Fahim Ferdous (Developer of Tkl)