nn-wtf aims at providing a convenience wrapper to Google's
[TensorFlow](http://www.tensorflow.org/) machine learning library.
Its focus is on making neural networks easy to set up, train and use.
The library is in pre-alpha right now and does not do anything seriously useful
yet.
## Installation
nn-wtf runs under Python3.4 and above.
### Dependencies
You need to install TensorFlow manually. The installation process depends on
your system. Install the version of TensorFlow built for Python 3.4.
See
https://www.tensorflow.org/versions/r0.8/get_started/os_setup.html#download-and-setup
for details.
Example installation for Linux without GPU support:
```
$ pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.8.0rc0-cp34-cp34m-linux_x86_64.whl
```
### NN-WTF itself
Simple:
```
$ pip install nn_wtf
```
## Documentation
Sorry the documentation is absolutely minimal at this point. More useful
documentation will be ready by the time this package reaches alpha status.
### List of useful classes and methods
* `NeuralNetworkGraph`: Base class for defining and training neural networks
* `__init__(self, input_size, layer_sizes, output_size, learning_rate)`
* `set_session(self, session=None)`
* `train(self, data_sets, max_steps, precision, steps_between_checks, run_as_check, batch_size)`
* `get_predictor().predict(input_data)`
* `MNISTGraph`: Subclass of NeuralNetworkGraph suitable for working on MNIST data
* `NeuralNetworkOptimizer`: Optimize geometry of a neural network for fast training
* `__init__( self, tested_network, input_size, output_size, training_precision,
layer_sizes, learning_rate, verbose, batch_size)`
* `brute_force_optimal_network_geometry(self, data_sets, max_steps)`
### Usage example
If you want to try it on MNIST data, try this sample program:
```python
from nn_wtf.mnist_data_sets import MNISTDataSets
from nn_wtf.mnist_graph import MNISTGraph
import tensorflow as tf
data_sets = MNISTDataSets('.')
graph = MNISTGraph(
learning_rate=0.1, layer_sizes=(64, 64, 16), train_dir='.'
)
graph.train(data_sets, max_steps=5000, precision=0.95)
image_data = MNISTDataSets.read_one_image_from_url(
'http://github.com/lene/nn-wtf/blob/master/nn_wtf/data/7_from_test_set.raw?raw=true'
)
prediction = graph.get_predictor().predict(image_data)
assert prediction == 7
```
>From there on, you are on your own for now. More functionality and documentation
to come.
Metadata
Release files for nn_wtf 0.1.7
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| nn_wtf-0.1.7.linux-x86_64.tar.gz | 81.9 kB | Without provenance | Details for nn_wtf-0.1.7.linux-x86_64.tar.gz |
Built distributions (wheels)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| nn_wtf-0.1.7-py3.5.egg | Legacy Egg format | - | - | Details for nn_wtf-0.1.7-py3.5.egg |
| nn_wtf-0.1.7-py2-none-any.whl | Python 2 | none | any | Details for nn_wtf-0.1.7-py2-none-any.whl |
Total release size: 240.1 kB
Release files / nn_wtf-0.1.7.linux-x86_64.tar.gz
| Download URL | nn_wtf-0.1.7.linux-x86_64.tar.gz |
|---|---|
| Size | 81.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
65741a9c8c40fa44bd2c356984729e1e22dbc502606ee4f945e0cae1cfb1db33
|
|
BLAKE2b-256 checksum How to use checksums |
ee0834b649abd41943a936b02648491f23f8bb8b8d74b7b76353681173e81c25
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Release files / nn_wtf-0.1.7-py3.5.egg
| Download URL | nn_wtf-0.1.7-py3.5.egg |
|---|---|
| Size | 105.5 kB |
| Tags | Egg |
|
SHA-256 checksum How to use checksums |
a4443d33d0776ce3b4e96bdc1c3937eb5a4fa0ab945917d58d57477a73bcada5
|
|
BLAKE2b-256 checksum How to use checksums |
ad81cbf10c9ac2559eb7eaf3c7611934c11ba0910f06828c0ce0454805be264c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Release files / nn_wtf-0.1.7-py2-none-any.whl
| Download URL | nn_wtf-0.1.7-py2-none-any.whl |
|---|---|
| Size | 52.7 kB |
| Tags | Python 2 |
|
SHA-256 checksum How to use checksums |
b40a7ef10a989ef1bda72bc2d0c4550c330fe60b87539fc395b420bd6ace8366
|
|
BLAKE2b-256 checksum How to use checksums |
abb6e5304b0580e2bbff398f5a24192c72ce435ed3371e8ea8334cf8acd9ba93
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Security for nn_wtf 0.1.7
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.