neka-nat/tensorboard-chainer

[![Build Status](https://travis-ci.org/neka-nat/tensorboard-chainer.svg?branch=master)](https://travis-ci.org/neka-nat/tensorboard-chainer)
[![codecov](https://codecov.io/gh/neka-nat/tensorboard-chainer/branch/master/graph/badge.svg)](https://codecov.io/gh/neka-nat/tensorboard-chainer)

[![Code Climate](https://codeclimate.com/github/neka-nat/tensorboard-chainer/badges/gpa.svg)](https://codeclimate.com/github/neka-nat/tensorboard-chainer)

[![PyPI version](https://badge.fury.io/py/tensorboard-chainer.svg)](https://badge.fury.io/py/tensorboard-chainer)

# tensorboard-chainer

Write tensorboard events with simple command.
including scalar, image, histogram, audio, text, graph and embedding.

This is based on [tensorboard-pytorch](https://github.com/lanpa/tensorboard-pytorch).

## Usage

Install tensorflow.

```
pip install tensorflow
```

Execute demo.py and tensorboard.
Access "localhost:6006" in your browser.

```
cd examples
python demo.py
tensorboard --logdir runs
```

## Scalar example

![graph](https://raw.githubusercontent.com/neka-nat/tensorboard-chainer/master/screenshots/scalar.png)

## Histogram example

![graph](https://raw.githubusercontent.com/neka-nat/tensorboard-chainer/master/screenshots/histogram.png)

## Graph example

![graph](https://raw.githubusercontent.com/neka-nat/tensorboard-chainer/master/screenshots/graph.gif)

## Name scope

Like tensorflow, nodes in the graph can be grouped together in the namespace to make it easy to see.

```python
import chainer
import chainer.functions as F
import chainer.links as L
from tb_chainer import name_scope, within_name_scope

class MLP(chainer.Chain):
    def __init__(self, n_units, n_out):
        super(MLP, self).__init__()
        with self.init_scope():
            self.l1 = L.Linear(None, n_units)  # n_in -> n_units
            self.l2 = L.Linear(None, n_units)  # n_units -> n_units
            self.l3 = L.Linear(None, n_out)  # n_units -> n_out

    @within_name_scope('MLP')
    def __call__(self, x):
        with name_scope('linear1', self.l1.params()):
            h1 = F.relu(self.l1(x))
        with name_scope('linear2', self.l2.params()):
            h2 = F.relu(self.l2(h1))
        with name_scope('linear3', self.l3.params()):
            o = self.l3(h2)
        return o
```

How to save the logs using this model is shown below.
`add_all_variable_images` is the function that saves the Variable's data in the model that matches the pattern as an images.
`add_all_parameter_histograms` is the function that save histograms of the Parameter's data in the model that match the pattern.

```python
from datetime import datetime
from tb_chainer import SummaryWriter

model = L.Classifier(MLP(1000, 10))

res = model(chainer.Variable(np.random.rand(1, 784).astype(np.float32)),
            chainer.Variable(np.random.rand(1).astype(np.int32)))

writer = SummaryWriter('runs/'+datetime.now().strftime('%B%d  %H:%M:%S'))
writer.add_graph([res])
writer.add_all_variable_images([res], pattern='.*MLP.*')
writer.add_all_parameter_histograms([res], pattern='.*MLP.*')

writer.close()
```

## Reference

* [tensorboard-pytorch](https://github.com/lanpa/tensorboard-pytorch)
* [tensorboard_logger](https://github.com/TeamHG-Memex/tensorboard_logger)
* [tfchain](https://github.com/mitmul/tfchain)