The Living Thing / Notebooks :


the framework to use for deep learning if you groupthink like Google

A C++/Python neural network toolkit by Google. I am using it for solving general machine-learning problems, and frequently enough that I have notes.


There are some other frontends, which seem a bit less useful to my mind:

I’m not convinced these latter options actually solve any problems. They seem to make the easy bits not easier but different, and the hard bits no easier.


See also keras tutorials below.


Google’s own Tensorflow without a phd.

Joonwook Choi recommends:

Basic ways:

Advanced ways:


Tensorboard is a de facto debugging tool standard. It’s not immediately intuitive; I recommend reading Li Yin’s explanation.


tensorboard --logdir=path/to/log-directory

or, more usually,

tensorboard --logdir=name1:/path/to/logs/1,name2:/path/to/logs/2 --host=localhost

or, lazily, (bash)

tensorboard --logdir=$(ls -dm *.logs |tr -d ' \n\r') --host=localhost


tensorboard --logdir=(string join , (for f in *.logs; echo (basename $f .logs):$f; end)) --host=localhost

In fact, that sometimes works not so well for me. Tensorboard reeeeally wants you to explicitly specify your folder names.

#!/bin/env python3
from pathlib import Path
from subprocess import run
import sys

p = Path('./')

logdirstring = '--logdir=' + ','.join([
    str(d)[:-5] + ":" + str(d)
    for d in p.glob('*.logs')

proc = run(

TensorBoard has a built-in visualizer, called the Embedding Projector, for interactive visualization and analysis of high-dimensional data like embeddings. It is meant to be useful for developers and researchers alike. It reads from the checkpoint files where you save your tensorflow variables. Although it’s most useful for embeddings, it will load any 2D tensor, potentially including your training weights.

Getting data in

This is a depressingly complex topic; Likely it’s more lines of code than building your actual learning algorithm.

For example, things break differently if

These interact in various different ways that seem irritating, but are probably to do with enabling very large scale data reading workflows, so that you might accidentally solve a problem for Google and they can get your solution for cheap.

Here’s a walk through of some of the details. And here are the manual pages for feeding and queueing

My experience that that stuff is so horribly messy that you should just build different graphs for the estimation and deployment phases of your mode and implement them each according to convenience. This of course is asking for trouble with errors

I’m not yet sure how to easily transmit the estimated parameters between graphs in these two separate phases… I’ll make notes about THAT when i come to it.

(Non-recurrent) convolutional networks

For the SAME padding, the output height and width are computed as:

python out_height = ceil(float(in_height) / float(strides[1])) out_width = ceil(float(in_width) / float(strides[2]))

For the VALID padding, the output height and width are computed as:

`python out_height = ceil(float(in_height - filter_height + 1) / float(strides[1])) out_width = ceil(float(in_width - filter_width + 1) / float(strides[2]))

[Tensorflow's 4d tensor packing for images](

> TensorFlow supports `NHWC` (default) and `NCHW` (cuDNN default).
> The best practice is to build models
> that work with both `NCHW` and `NHWC`
> as it is common to train using NCHW on GPU, and then do inference with NHWC on CPU.

`NCHW` is, to be clear, `(batch, channels, height, width)`.

Theano by contrast, is AFAICT always `NCHW`.

## Recurrent/fancy networks
The documentation for these is abysmal.

To write: How to create standard
[linear filters]({filename} in Tensorflow.

For now, my recommendation is to simply use [keras](, which makes this easier
inside tensorflow, or [pytorch]({filename}, which makes it easier overall.

[tensorflow fold]( is a library which ingests structured data and simulates
[pytorch]({filename} dynamic graphs dependent upon its structure.

### Official documentation
The Tensorflow RNN documentation, as bad as it is, is not even easy to find,
being scattered across several non-obvious locations without consistent

* [Overview docs](

* [Other docs of confusing relation to the prior docs](

* [tutorial docs](

* stateful minibatch training requires the catchy [SequenceQueueingStateSaver](

To make it actually make sense without unwarranted time wasting and guessing,
you will then need to read other stuff.

### Community guides
* `seq2seq` models with GRUs : [Fun with Recurrent Neural Nets](

* [Variable sequence length HOWTO](

* [Where do the RNN weights come from](

* Denny Britz's blog posts
  * [RNNs in Tensorflow, a practical guide and undocumented features](
  * He also gives a good explanation of [vanishing gradients](

* Danijar Hafner
  * [Introduction to Recurrent Networks in TensorFlow](
  * [Variable sequence lengths HOWTO](

* Philippe Remy, [Stateful LSTM in Keras](

* Ben Bolte, [Deep Language Modeling for Question Answering using Keras](

* pro tip: [SequenceQueueingStateSaver]( makes things easy.

## Keras: The recommended way of using tensorflow
You probably want to start using a higher level `keras`
unless your needs are extraordinarily esoteric or
you like reinventing wheels.
Keras is a good choice, since it removes a lot of boilerplate,
and makes even writing new boilerplate easier.

It adds only a few minor restrictions to your abilities,
but by creating a consistent API, has become something of a standard for early
access to complex new algorithms you would never have time to re-implement

I would use it if I were you for anything involving standard neural networks,
especially any kind of recurrent network.
If you want to optimise a *generic*, non-deep neural model,
you might find the naked tensorflow API has less friction.

* [Easing pain via Keras](

* Jason Brownlee's [HOWTO guide](

* [Recurrent neural networks' gradients are truncated to the sequence length](,
  which might not be
  obvious. But this is the TBPTT parameter.

* [recurrentshop]( makes it easier to manage recurrent topologies using [keras](

## Getting models out
* For a local app:
  Hamed MP, [Exporting trained TensorFlow models to C++ the RIGHT way!](

* For serving it online, Tensorflow `serving` is the preferred method.
  See [the Serving documentation](

* for mobile app the [HBO joke hotdog app HOWTO]( gives a wonderful explanation.

## Training in the cloud because you don't have NVIDIA sponsorship
See [practical cloud computing]({filename},
which has a couple of sections on that.

## Extending
Tensorflow [allows binary extensions](
but don't really explain how it integrates with normal python builds.
Here is [an example from Uber](

## Misc HOWTOs
### Nightly builds
(or build your own)

### Dynamic graphs
[Pytorch]({filename} has JIT graphs and they are super hip, so now tensorflow has a
[dynamic graph mode](,
called `Eager`.

### GPU selection
sets `NVIDIA_VISIBLE_GPU` to the least loaded GPU.

### Silencing tensorflow
TF_CPP_MIN_LOG_LEVEL=1 primusrun python biquad_fast

Hessians and higher order optimisation

Basic Newton method optimisation example. Very basic example that also shows how to create a diagonal hessian.

Slightly outdated, Hessian matrix. There is a discussion on Jacobians in TF, including, e.g. some fancy examples by jjough:

here’s mine — works for high-dimensional Jacobians (numerator and denominator have >1 dimension), undefined batch sizes, and tensors that are not statically known.

Remember to use an interactive session, otherwise tf.get_default_session() will not be able to find the session.

python def tf_jacobian(tensor2, tensor1, feed_dict, sess = tf.get_default_session()): """ Computes the tensor d(tensor2)/d(tensor1) recursively. :param tensor2: numerator of Jacobian :param tensor1: denominator of Jacobian :param feed_dict: input data (need this if tensors are not statically known) :return: a tensor of dimension (dim_tensor2 x dim_tensor1) """ # can't do tensor.get_shape() because it doesn't work for undefined batch size shape = list(, feed_dict)) if shape: # split tensor2 along first dimension and recur # int trick from tensor2_split = tf.split(axis = 0, num_or_size_splits = int(shape[0]), value = tensor2) grad_split = [tf_jacobian(tf.squeeze(M, squeeze_dims = 0), tensor1, feed_dict) for M in tensor2_split] return tf.stack(grad_split) else: # calculate gradient of scalar grad = tf.gradients(tensor2, tensor1) if grad[0] != None: return tf.squeeze(grad, squeeze_dims = [0]) else: # replace any undefined gradients with zeros return tf.zeros_like(tensor1)

And here’s one for batched tensors:

`python def batch_tf_jacobian(tensor2, tensor1, feed_dict, sess = tf.get_default_session()): “”” Computes the matrix d(tensor2)/d(tensor1) recursively. Tensorflow doesn’t really have its own Jacobian operator (tf.gradients sums over all dims of tensor2).

:param tensor2: numerator of Jacobian, first dimension is batch
:param tensor1: denominator of Jacobian, first dimension is batch
:param feed_dict: input data (need this if tensors are not statically known)
:return: batch Jacobian tensor
shape2 = list(, feed_dict))
shape1 = list(, feed_dict))

jacobian = tf_jacobian(tensor2, tensor1, feed_dict)
batch_size = shape2[0]

batch_jacobian = [tf.slice(jacobian, [i] + [0]*(len(shape2)-1) + [i] + [0]*(len(shape1)-1),  [1] + [-1]*(len(shape2)-1) + [1] + [-1]*(len(shape1)-1)) for i in range(batch_size)]
batch_jacobian = [tf.squeeze(tensor, squeeze_dims = (0, len(shape2))) for tensor in batch_jacobian]
batch_jacobian = tf.stack(batch_jacobian)
return batch_jacobian


Manage tensorflow environments


Optimisation tricks

Using traditional/experimental optimisers rather than SGD-type ones.

Simplify distributed training using Horovod.