TensorFlow: Upstream latest commits to git.
Changes: - Updates to Documentation, README.md, installation instructions, anchor links, etc. - Adds Readme for embedding directory. Base CL: 107308461
This commit is contained in:
Vijay Vasudevan
parent
8bd3b38e66
commit
7de9099a73
17
README.md
17
README.md
@ -11,9 +11,21 @@ organization for the purposes of conducting machine learning and deep neural
|
||||
networks research. The system is general enough to be applicable in a wide
|
||||
variety of other domains, as well.
|
||||
|
||||
|
||||
**Note: Currently we do not accept pull requests on github -- see
|
||||
[CONTRIBUTING.md](CONTRIBUTING.md) for information on how to contribute code
|
||||
changes to TensorFlow through
|
||||
[tensorflow.googlesource.com](https://tensorflow.googlesource.com/tensorflow)**
|
||||
|
||||
**We use [github issues](https://github.com/tensorflow/tensorflow/issues) for
|
||||
tracking requests and bugs, but please see
|
||||
[Community](resources/index.md#community) for general questions and
|
||||
discussion.**
|
||||
|
||||
# Download and Setup
|
||||
|
||||
For detailed installation instructions, see
|
||||
To install TensorFlow using a binary package, see the instructions below. For
|
||||
more detailed installation instructions, including installing from source, see
|
||||
[here](tensorflow/g3doc/get_started/os_setup.md).
|
||||
|
||||
## Binary Installation
|
||||
@ -32,7 +44,8 @@ Install TensorFlow:
|
||||
# For CPU-only version
|
||||
$ sudo pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl
|
||||
|
||||
# For GPU-enabled version
|
||||
# For GPU-enabled version. See detailed install instructions
|
||||
# for GPU configuration information.
|
||||
$ sudo pip install https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl
|
||||
```
|
||||
|
||||
|
||||
@ -5,6 +5,7 @@
|
||||
## Contents
|
||||
* [Session management](#AUTOGENERATED-session-management)
|
||||
* [class tf.Session](#Session)
|
||||
* [class tf.InteractiveSession](#InteractiveSession)
|
||||
* [tf.get_default_session()](#get_default_session)
|
||||
* [Error classes](#AUTOGENERATED-error-classes)
|
||||
* [class tf.OpError](#OpError)
|
||||
@ -257,6 +258,78 @@ thread's function.
|
||||
|
||||
|
||||
|
||||
- - -
|
||||
|
||||
### class tf.InteractiveSession <div class="md-anchor" id="InteractiveSession">{#InteractiveSession}</div>
|
||||
|
||||
A TensorFlow `Session` for use in interactive contexts, such as a shell.
|
||||
|
||||
The only difference with a regular `Session` is that an `InteractiveSession`
|
||||
installs itself as the default session on construction.
|
||||
The methods [`Tensor.eval()`](framework.md#Tensor.eval) and
|
||||
[`Operation.run()`](framework.md#Operation.run) will use that session
|
||||
to run ops.
|
||||
|
||||
This is convenient in interactive shells and [IPython
|
||||
notebooks](http://ipython.org), as it avoids having to pass an explicit
|
||||
`Session` object to run ops.
|
||||
|
||||
For example:
|
||||
|
||||
```python
|
||||
sess = tf.InteractiveSession()
|
||||
a = tf.constant(5.0)
|
||||
b = tf.constant(6.0)
|
||||
c = a * b
|
||||
# We can just use 'c.eval()' without passing 'sess'
|
||||
print c.eval()
|
||||
sess.close()
|
||||
```
|
||||
|
||||
Note that a regular session installs itself as the default session when it
|
||||
is created in a `with` statement. The common usage in non-interactive
|
||||
programs is to follow that pattern:
|
||||
|
||||
```python
|
||||
a = tf.constant(5.0)
|
||||
b = tf.constant(6.0)
|
||||
c = a * b
|
||||
with tf.Session():
|
||||
# We can also use 'c.eval()' here.
|
||||
print c.eval()
|
||||
```
|
||||
|
||||
- - -
|
||||
|
||||
#### tf.InteractiveSession.__init__(target='', graph=None) {#InteractiveSession.__init__}
|
||||
|
||||
Creates a new interactive TensorFlow session.
|
||||
|
||||
If no `graph` argument is specified when constructing the session,
|
||||
the default graph will be launched in the session. If you are
|
||||
using more than one graph (created with `tf.Graph()` in the same
|
||||
process, you will have to use different sessions for each graph,
|
||||
but each graph can be used in multiple sessions. In this case, it
|
||||
is often clearer to pass the graph to be launched explicitly to
|
||||
the session constructor.
|
||||
|
||||
##### Args:
|
||||
|
||||
|
||||
* <b>target</b>: (Optional.) The execution engine to connect to.
|
||||
Defaults to using an in-process engine. At present, no value
|
||||
other than the empty string is supported.
|
||||
* <b>graph</b>: (Optional.) The `Graph` to be launched (described above).
|
||||
|
||||
|
||||
- - -
|
||||
|
||||
#### tf.InteractiveSession.close() {#InteractiveSession.close}
|
||||
|
||||
Closes an `InteractiveSession`.
|
||||
|
||||
|
||||
|
||||
|
||||
- - -
|
||||
|
||||
|
||||
@ -299,6 +299,7 @@
|
||||
* [`DeadlineExceededError`](client.md#DeadlineExceededError)
|
||||
* [`FailedPreconditionError`](client.md#FailedPreconditionError)
|
||||
* [`get_default_session`](client.md#get_default_session)
|
||||
* [`InteractiveSession`](client.md#InteractiveSession)
|
||||
* [`InternalError`](client.md#InternalError)
|
||||
* [`InvalidArgumentError`](client.md#InvalidArgumentError)
|
||||
* [`NotFoundError`](client.md#NotFoundError)
|
||||
|
||||
@ -54,10 +54,10 @@ suggest skimming blue, then red.
|
||||
|
||||
<div style="width:100%; margin:auto; margin-bottom:10px; margin-top:20px; display: flex; flex-direction: row">
|
||||
<a href="../tutorials/mnist/beginners/index.md">
|
||||
<img style="flex-grow:1; flex-shrink:1;border: 1px solid black;" src="./blue_pill.png">
|
||||
<img style="flex-grow:1; flex-shrink:1; border: 1px solid black;" src="blue_pill.png">
|
||||
</a>
|
||||
<a href="../tutorials/mnist/pros/index.md">
|
||||
<img style="flex-grow:1; flex-shrink:1; border: 1px solid black;" src="./red_pill.png">
|
||||
<img style="flex-grow:1; flex-shrink:1; border: 1px solid black;" src="red_pill.png">
|
||||
</a>
|
||||
</div>
|
||||
<p style="font-size:10px;">Images licensed CC BY-SA 4.0; original by W. Carter</p>
|
||||
|
||||
@ -88,35 +88,8 @@ ImportError: libcudart.so.7.0: cannot open shared object file: No such file or d
|
||||
you most likely need to set your `LD_LIBRARY_PATH` to point to the location of
|
||||
your CUDA libraries.
|
||||
|
||||
### Train the MNIST neural net model
|
||||
|
||||
```sh
|
||||
$ python tensorflow/models/image/mnist/convolutional.py
|
||||
Succesfully downloaded train-images-idx3-ubyte.gz 9912422 bytes.
|
||||
Succesfully downloaded train-labels-idx1-ubyte.gz 28881 bytes.
|
||||
Succesfully downloaded t10k-images-idx3-ubyte.gz 1648877 bytes.
|
||||
Succesfully downloaded t10k-labels-idx1-ubyte.gz 4542 bytes.
|
||||
Extracting data/train-images-idx3-ubyte.gz
|
||||
Extracting data/train-labels-idx1-ubyte.gz
|
||||
Extracting data/t10k-images-idx3-ubyte.gz
|
||||
Extracting data/t10k-labels-idx1-ubyte.gz
|
||||
can't determine number of CPU cores: assuming 4
|
||||
I tensorflow/core/common_runtime/local_device.cc:25] Local device intra op
|
||||
parallelism threads: 3
|
||||
can't determine number of CPU cores: assuming 4
|
||||
I tensorflow/core/common_runtime/local_session.cc:45] Local session inter op
|
||||
parallelism threads: 4
|
||||
Initialized!
|
||||
Epoch 0.00
|
||||
Minibatch loss: 12.054, learning rate: 0.010000
|
||||
Minibatch error: 90.6%
|
||||
Validation error: 84.6%
|
||||
...
|
||||
...
|
||||
|
||||
```
|
||||
|
||||
## Installing from sources {#source}
|
||||
<a name="source"></a>
|
||||
## Installing from sources
|
||||
|
||||
### Clone the TensorFlow repository
|
||||
|
||||
@ -260,7 +233,8 @@ Notes : You need to install
|
||||
Follow installation instructions [here](http://docs.scipy.org/doc/numpy/user/install.html).
|
||||
|
||||
|
||||
### Create the pip package and install {#create-pip}
|
||||
<a name="create-pip"></a>
|
||||
### Create the pip package and install
|
||||
|
||||
```sh
|
||||
$ bazel build -c opt //tensorflow/tools/pip_package:build_pip_package
|
||||
|
||||
@ -1,4 +1,4 @@
|
||||
# Adding a New Op to TensorFlow
|
||||
# Adding a New Op
|
||||
|
||||
PREREQUISITES:
|
||||
|
||||
@ -27,27 +27,28 @@ to:
|
||||
|
||||
<!-- TOC-BEGIN This section is generated by neural network: DO NOT EDIT! -->
|
||||
## Contents
|
||||
* [Define the Op's interface](#define_interface)
|
||||
* [Define the Op's interface](#AUTOGENERATED-define-the-op-s-interface)
|
||||
* [Implement the kernel for the Op](#AUTOGENERATED-implement-the-kernel-for-the-op)
|
||||
* [Generate the client wrapper](#AUTOGENERATED-generate-the-client-wrapper)
|
||||
* [The Python Op wrapper](#AUTOGENERATED-the-python-op-wrapper)
|
||||
* [The C++ Op wrapper](#AUTOGENERATED-the-c---op-wrapper)
|
||||
* [Verify it works](#AUTOGENERATED-verify-it-works)
|
||||
* [Validation](#validation)
|
||||
* [Validation](#AUTOGENERATED-validation)
|
||||
* [Op registration](#AUTOGENERATED-op-registration)
|
||||
* [Attrs](#AUTOGENERATED-attrs)
|
||||
* [Attr types](#AUTOGENERATED-attr-types)
|
||||
* [Polymorphism](#polymorphism)
|
||||
* [Polymorphism](#AUTOGENERATED-polymorphism)
|
||||
* [Inputs and Outputs](#AUTOGENERATED-inputs-and-outputs)
|
||||
* [Backwards compatibility](#AUTOGENERATED-backwards-compatibility)
|
||||
* [GPU Support](#mult-archs)
|
||||
* [GPU Support](#AUTOGENERATED-gpu-support)
|
||||
* [Implement the gradient in Python](#AUTOGENERATED-implement-the-gradient-in-python)
|
||||
* [Implement a shape function in Python](#AUTOGENERATED-implement-a-shape-function-in-python)
|
||||
|
||||
|
||||
<!-- TOC-END This section was generated by neural network, THANKS FOR READING! -->
|
||||
|
||||
## Define the Op's interface <div class="md-anchor" id="define_interface">{#define_interface}</div>
|
||||
<a name="define_interface"></a>
|
||||
## Define the Op's interface <div class="md-anchor" id="AUTOGENERATED-define-the-op-s-interface">{#AUTOGENERATED-define-the-op-s-interface}</div>
|
||||
|
||||
You define the interface of an Op by registering it with the TensorFlow system.
|
||||
In the registration, you specify the name of your Op, its inputs (types and
|
||||
@ -210,7 +211,7 @@ Then run your test:
|
||||
$ bazel test tensorflow/python:zero_out_op_test
|
||||
```
|
||||
|
||||
## Validation <div class="md-anchor" id="validation">{#validation}</div>
|
||||
## Validation <div class="md-anchor" id="AUTOGENERATED-validation">{#AUTOGENERATED-validation}</div>
|
||||
|
||||
The example above assumed that the Op applied to a tensor of any shape. What
|
||||
if it only applied to vectors? That means adding a check to the above OpKernel
|
||||
@ -445,8 +446,8 @@ REGISTER_OP("AttrDefaultExampleForAllTypes")
|
||||
Note in particular that the values of type `type` use [the `DT_*` names
|
||||
for the types](../../resources/dims_types.md#data-types).
|
||||
|
||||
### Polymorphism <div class="md-anchor" id="polymorphism">{#polymorphism}</div>
|
||||
#### Type Polymorphism {#type-polymorphism}
|
||||
### Polymorphism <div class="md-anchor" id="AUTOGENERATED-polymorphism">{#AUTOGENERATED-polymorphism}</div>
|
||||
#### Type Polymorphism
|
||||
|
||||
For ops that can take different types as input or produce different output
|
||||
types, you can specify [an attr](#attrs) in
|
||||
@ -466,7 +467,8 @@ REGISTER\_OP("ZeroOut")
|
||||
Your Op registration now specifies that the input's type must be `float`, or
|
||||
`int32`, and that its output will be the same type, since both have type `T`.
|
||||
|
||||
> A note on naming:{#naming} Inputs, outputs, and attrs generally should be
|
||||
<a name="naming"></a>
|
||||
> A note on naming: Inputs, outputs, and attrs generally should be
|
||||
> given snake_case names. The one exception is attrs that are used as the type
|
||||
> of an input or in the type of an input. Those attrs can be inferred when the
|
||||
> op is added to the graph and so don't appear in the op's function. For
|
||||
@ -482,7 +484,7 @@ Your Op registration now specifies that the input's type must be `float`, or
|
||||
> name: A name for the operation (optional).
|
||||
>
|
||||
> Returns:
|
||||
> A `Tensor`. Has the same type as `x`.
|
||||
> A `Tensor`. Has the same type as `to_zero`.
|
||||
> """
|
||||
> ```
|
||||
>
|
||||
@ -674,7 +676,8 @@ TF_CALL_REAL_NUMBER_TYPES(REGISTER_KERNEL);
|
||||
#undef REGISTER_KERNEL
|
||||
```
|
||||
|
||||
#### List Inputs and Outputs {#list-input-output}
|
||||
<a name="list-input-output"></a>
|
||||
#### List Inputs and Outputs
|
||||
|
||||
In addition to being able to accept or produce different types, ops can consume
|
||||
or produce a variable number of tensors.
|
||||
@ -894,7 +897,8 @@ There are several ways to preserve backwards-compatibility.
|
||||
If you cannot make your change to an operation backwards compatible, then
|
||||
create a new operation with a new name with the new semantics.
|
||||
|
||||
## GPU Support <div class="md-anchor" id="mult-archs">{#mult-archs}</div>
|
||||
<a name="mult-archs"></a>
|
||||
## GPU Support <div class="md-anchor" id="AUTOGENERATED-gpu-support">{#AUTOGENERATED-gpu-support}</div>
|
||||
|
||||
You can implement different OpKernels and register one for CPU and another for
|
||||
GPU, just like you can [register kernels for different types](#polymorphism).
|
||||
|
||||
@ -1,4 +1,4 @@
|
||||
# TensorBoard: Visualizing Your Graph
|
||||
# TensorBoard: Graph Visualization
|
||||
|
||||
TensorFlow computation graphs are powerful but complicated. The graph visualization can help you understand and debug them. Here's an example of the visualization at work.
|
||||
|
||||
|
||||
@ -18,7 +18,7 @@ example.
|
||||
[View Tutorial](../tutorials/mnist/tf/index.md)
|
||||
|
||||
|
||||
## TensorBoard: Visualizing Your Training
|
||||
## TensorBoard: Visualizing Learning
|
||||
|
||||
TensorBoard is a useful tool for visualizing the training and evaluation of
|
||||
your model(s). This tutorial describes how to build and run TensorBoard as well
|
||||
@ -28,7 +28,7 @@ TensorBoard uses for display.
|
||||
[View Tutorial](summaries_and_tensorboard/index.md)
|
||||
|
||||
|
||||
## TensorBoard: Visualizing Your Graph
|
||||
## TensorBoard: Graph Visualization
|
||||
|
||||
This tutorial describes how to use the graph visualizer in TensorBoard to help
|
||||
you understand the dataflow graph and debug it.
|
||||
@ -60,7 +60,7 @@ compose in your graph, but here are the details of how to add you own custom Op.
|
||||
[View Tutorial](adding_an_op/index.md)
|
||||
|
||||
|
||||
## New Data Formats
|
||||
## Custom Data Readers
|
||||
|
||||
If you have a sizable custom data set, you may want to consider extending
|
||||
TensorFlow to read your data directly in it's native format. Here's how.
|
||||
@ -68,7 +68,7 @@ TensorFlow to read your data directly in it's native format. Here's how.
|
||||
[View Tutorial](new_data_formats/index.md)
|
||||
|
||||
|
||||
## Using One or More GPUs
|
||||
## Using GPUs
|
||||
|
||||
This tutorial describes how to construct and execute models on GPU(s).
|
||||
|
||||
|
||||
@ -1,4 +1,4 @@
|
||||
# Extending TF: Supporting new data formats
|
||||
# Custom Data Readers
|
||||
|
||||
PREREQUISITES:
|
||||
|
||||
|
||||
@ -10,13 +10,13 @@ There are three main methods of getting data into a TensorFlow program:
|
||||
|
||||
<!-- TOC-BEGIN This section is generated by neural network: DO NOT EDIT! -->
|
||||
## Contents
|
||||
* [Feeding](#Feeding)
|
||||
* [Feeding](#AUTOGENERATED-feeding)
|
||||
* [Reading from files](#AUTOGENERATED-reading-from-files)
|
||||
* [Filenames, shuffling, and epoch limits](#AUTOGENERATED-filenames--shuffling--and-epoch-limits)
|
||||
* [File formats](#AUTOGENERATED-file-formats)
|
||||
* [Preprocessing](#AUTOGENERATED-preprocessing)
|
||||
* [Batching](#AUTOGENERATED-batching)
|
||||
* [Creating threads to prefetch using `QueueRunner` objects](#QueueRunner)
|
||||
* [Creating threads to prefetch using `QueueRunner` objects](#AUTOGENERATED-creating-threads-to-prefetch-using--queuerunner--objects)
|
||||
* [Filtering records or producing multiple examples per record](#AUTOGENERATED-filtering-records-or-producing-multiple-examples-per-record)
|
||||
* [Sparse input data](#AUTOGENERATED-sparse-input-data)
|
||||
* [Preloaded data](#AUTOGENERATED-preloaded-data)
|
||||
@ -25,7 +25,7 @@ There are three main methods of getting data into a TensorFlow program:
|
||||
|
||||
<!-- TOC-END This section was generated by neural network, THANKS FOR READING! -->
|
||||
|
||||
## Feeding <div class="md-anchor" id="Feeding">{#Feeding}</div>
|
||||
## Feeding <div class="md-anchor" id="AUTOGENERATED-feeding">{#AUTOGENERATED-feeding}</div>
|
||||
|
||||
TensorFlow's feed mechanism lets you inject data into any Tensor in a
|
||||
computation graph. A python computation can thus feed data directly into the
|
||||
@ -267,7 +267,8 @@ summary to the graph that indicates how full the example queue is. If you have
|
||||
enough reading threads, that summary will stay above zero. You can
|
||||
[view your summaries as training progresses using TensorBoard](../summaries_and_tensorboard/index.md).
|
||||
|
||||
### Creating threads to prefetch using `QueueRunner` objects <div class="md-anchor" id="QueueRunner">{#QueueRunner}</div>
|
||||
<a name="QueueRunner"></a>
|
||||
### Creating threads to prefetch using `QueueRunner` objects <div class="md-anchor" id="AUTOGENERATED-creating-threads-to-prefetch-using--queuerunner--objects">{#AUTOGENERATED-creating-threads-to-prefetch-using--queuerunner--objects}</div>
|
||||
|
||||
The short version: many of the `tf.train` functions listed above add
|
||||
[`QueueRunner`](../../api_docs/python/train.md#QueueRunner) objects to your
|
||||
|
||||
@ -1,4 +1,4 @@
|
||||
# TensorBoard: Visualizing Your Training
|
||||
# TensorBoard: Visualizing Learning
|
||||
|
||||
The computations you'll use TensorBoard for - like training a massive
|
||||
deep neural network - can be complex and confusing. To make it easier to
|
||||
|
||||
@ -11,8 +11,8 @@ implementation can be found in out white paper:
|
||||
### Citation
|
||||
|
||||
If you use TensorFlow in your research and would like to cite the TensorFlow
|
||||
system, we suggest you cite the paper above. You can use this [BibTeX
|
||||
entry](bib.md). As the project progresses, we
|
||||
system, we suggest you cite the paper above.
|
||||
You can use this [BibTeX entry](bib.md). As the project progresses, we
|
||||
may update the suggested citation with new papers.
|
||||
|
||||
|
||||
|
||||
@ -33,11 +33,11 @@ experiments.
|
||||
The CIFAR-10 tutorial demonstrates several important constructs for
|
||||
designing larger and more sophisticated models in TensorFlow:
|
||||
|
||||
* Core mathematical components including
|
||||
[convolution](../../api_docs/python/nn.md#conv2d),
|
||||
[rectified linear activations](../../api_docs/python/nn.md#relu),
|
||||
[max pooling](../../api_docs/python/nn.md#max_pool) and
|
||||
[local response normalization](../../api_docs/python/nn.md#local_response_normalization).
|
||||
* Core mathematical components including[convolution](
|
||||
../../api_docs/python/nn.md#conv2d), [rectified linear activations](
|
||||
../../api_docs/python/nn.md#relu), [max pooling](
|
||||
../../api_docs/python/nn.md#max_pool) and [local response normalization](
|
||||
../../api_docs/python/nn.md#local_response_normalization).
|
||||
* [Visualization](../../how_tos/summaries_and_tensorboard/index.md)
|
||||
of network activity during training including input images,
|
||||
losses and distributions of activations and gradients.
|
||||
@ -341,8 +341,8 @@ all learned model parameters with the moving average version. This
|
||||
substitution boosts model performance at evaluation time.
|
||||
|
||||
> **EXERCISE:** Employing averaged parameters may boost predictive performance
|
||||
by about 3% as measured by precision@1. Edit `cifar10_eval.py` to not employ the
|
||||
averaged parameters for the model and verify that the predictive performance
|
||||
by about 3% as measured by precision @ 1. Edit `cifar10_eval.py` to not employ
|
||||
the averaged parameters for the model and verify that the predictive performance
|
||||
drops.
|
||||
|
||||
|
||||
|
||||
@ -230,7 +230,8 @@ with encoder inputs representing `[PAD PAD "." "go" "I"]` and decoder
|
||||
inputs `[GO "Je" "vais" "." EOS PAD PAD PAD PAD PAD]`.
|
||||
|
||||
|
||||
## Let's Run It {#run_it}
|
||||
<a name="run_it"></a>
|
||||
## Let's Run It
|
||||
|
||||
To train the model described above, we need to a large English-French corpus.
|
||||
We will use the *10^9-French-English corpus* from the
|
||||
|
||||
53
tensorflow/models/embedding/README.md
Normal file
53
tensorflow/models/embedding/README.md
Normal file
@ -0,0 +1,53 @@
|
||||
This directory contains models for unsupervised training of word embeddings
|
||||
using the model described in:
|
||||
|
||||
(Mikolov, et. al.) [Efficient Estimation of Word Representations in Vector Space](http://arxiv.org/abs/1301.3781),
|
||||
ICLR 2013.
|
||||
|
||||
Detailed instructions on how to get started and use them are available in the
|
||||
tutorials. Brief instructions are below.
|
||||
|
||||
* [Word2Vec Tutorial](http://tensorflow.org/tutorials/word2vec/)
|
||||
|
||||
To download the example text and evaluation data:
|
||||
|
||||
```shell
|
||||
wget http://mattmahoney.net/dc/text8.zip -O text8.gz
|
||||
gzip -d text8.gz -f
|
||||
wget https://word2vec.googlecode.com/svn/trunk/questions-words.txt
|
||||
```
|
||||
|
||||
To build everything under models/embedding/...:
|
||||
|
||||
```shell
|
||||
bazel build -c opt tensorflow/models/embedding:all
|
||||
```
|
||||
|
||||
To run the code manually:
|
||||
|
||||
```shell
|
||||
bazel-bin/tensorflow/models/embedding/word2vec_optimized \
|
||||
--train_data=text8 \
|
||||
--eval_data=questions-words.txt \
|
||||
--save_path=/tmp/
|
||||
```
|
||||
|
||||
To run the code via bazel:
|
||||
|
||||
```shell
|
||||
bazel run -c opt tensorflow/models/embedding/word2vec_optimized -- \
|
||||
--train_data=text8 \
|
||||
--eval_data=questions-words.txt \
|
||||
--save_path=/tmp/
|
||||
```
|
||||
|
||||
Here is a short overview of what is in this directory.
|
||||
|
||||
File | What's in it?
|
||||
--- | ---
|
||||
`word2vec.py` | A version of word2vec implemented using Tensorflow ops and minibatching.
|
||||
`word2vec_test.py` | Integration test for word2vec.
|
||||
`word2vec_optimized.py` | A version of word2vec implemented using C ops that does no minibatching.
|
||||
`word2vec_optimized_test.py` | Integration test for word2vec_optimized.
|
||||
`word2vec_kernels.cc` | Kernels for the custom input and training ops.
|
||||
`word2vec_ops.cc` | The declarations of the custom ops.
|
||||
@ -14,6 +14,7 @@ The key ops used are:
|
||||
* skipgram custom op that does input processing.
|
||||
"""
|
||||
|
||||
import os
|
||||
import sys
|
||||
import threading
|
||||
import time
|
||||
@ -357,7 +358,7 @@ class Word2Vec(object):
|
||||
def save_vocab(self):
|
||||
"""Save the vocabulary to a file so the model can be reloaded."""
|
||||
opts = self._options
|
||||
with open(opts.save_path + "/vocab.txt", "w") as f:
|
||||
with open(os.path.join(opts.save_path, "vocab.txt"), "w") as f:
|
||||
for i in xrange(opts.vocab_size):
|
||||
f.write(opts.vocab_words[i] + " " + str(opts.vocab_counts[i]) + "\n")
|
||||
|
||||
@ -493,7 +494,7 @@ def main(_):
|
||||
model.eval() # Eval analogies.
|
||||
# Perform a final save.
|
||||
model.saver.save(session,
|
||||
opts.save_path + "model",
|
||||
os.path.join(opts.save_path, "model.ckpt"),
|
||||
global_step=model.global_step)
|
||||
if FLAGS.interactive:
|
||||
# E.g.,
|
||||
|
||||
@ -13,6 +13,7 @@ The key ops used are:
|
||||
true SGD.
|
||||
"""
|
||||
|
||||
import os
|
||||
import sys
|
||||
import threading
|
||||
import time
|
||||
@ -217,7 +218,7 @@ class Word2Vec(object):
|
||||
def save_vocab(self):
|
||||
"""Save the vocabulary to a file so the model can be reloaded."""
|
||||
opts = self._options
|
||||
with open(opts.save_path + "/vocab.txt", "w") as f:
|
||||
with open(os.path.join(opts.save_path, "vocab.txt"), "w") as f:
|
||||
for i in xrange(opts.vocab_size):
|
||||
f.write(opts.vocab_words[i] + " " + str(opts.vocab_counts[i]) + "\n")
|
||||
|
||||
@ -396,7 +397,8 @@ def main(_):
|
||||
model.train() # Process one epoch
|
||||
model.eval() # Eval analogies.
|
||||
# Perform a final save.
|
||||
model.saver.save(session, opts.save_path + "model", global_step=model.step)
|
||||
model.saver.save(session, os.path.join(opts.save_path, "model.ckpt"),
|
||||
global_step=model.step)
|
||||
if FLAGS.interactive:
|
||||
# E.g.,
|
||||
# [0]: model.Analogy('france', 'paris', 'russia')
|
||||
|
||||
@ -7,6 +7,7 @@ examples of how a graph is launched in a [`tf.Session`](#Session).
|
||||
## Session management
|
||||
|
||||
@@Session
|
||||
@@InteractiveSession
|
||||
|
||||
@@get_default_session
|
||||
|
||||
|
||||
@ -524,32 +524,61 @@ class Session(BaseSession):
|
||||
class InteractiveSession(BaseSession):
|
||||
"""A TensorFlow `Session` for use in interactive contexts, such as a shell.
|
||||
|
||||
In some cases, such as interactive shells and IPython notebooks, it is
|
||||
useful to be able to define a `Session` without using a with block: this
|
||||
style enables statements to be executed immediately, rather than at the
|
||||
termination of the block. In that case, it must be closed using
|
||||
`Session.close()`. For example:
|
||||
The only difference with a regular `Session` is that an `InteractiveSession`
|
||||
installs itself as the default session on construction.
|
||||
The methods [`Tensor.eval()`](framework.md#Tensor.eval) and
|
||||
[`Operation.run()`](framework.md#Operation.run) will use that session
|
||||
to run ops.
|
||||
|
||||
This is convenient in interactive shells and [IPython
|
||||
notebooks](http://ipython.org), as it avoids having to pass an explicit
|
||||
`Session` object to run ops.
|
||||
|
||||
For example:
|
||||
|
||||
```python
|
||||
sess = InteractiveSession()
|
||||
sess = tf.InteractiveSession()
|
||||
a = tf.constant(5.0)
|
||||
b = tf.constant(6.0)
|
||||
c = a * b
|
||||
print c.run()
|
||||
# We can just use 'c.eval()' without passing 'sess'
|
||||
print c.eval()
|
||||
sess.close()
|
||||
```
|
||||
|
||||
Note that a regular session installs itself as the default session when it
|
||||
is created in a `with` statement. The common usage in non-interactive
|
||||
programs is to follow that pattern:
|
||||
|
||||
```python
|
||||
a = tf.constant(5.0)
|
||||
b = tf.constant(6.0)
|
||||
c = a * b
|
||||
with tf.Session():
|
||||
# We can also use 'c.eval()' here.
|
||||
print c.eval()
|
||||
```
|
||||
|
||||
@@__init__
|
||||
@@close
|
||||
"""
|
||||
|
||||
def __init__(self, target='', graph=None):
|
||||
"""Initializes an `InteractiveSession` object similar to `Session`.
|
||||
"""Creates a new interactive TensorFlow session.
|
||||
|
||||
If no `graph` argument is specified when constructing the session,
|
||||
the default graph will be launched in the session. If you are
|
||||
using more than one graph (created with `tf.Graph()` in the same
|
||||
process, you will have to use different sessions for each graph,
|
||||
but each graph can be used in multiple sessions. In this case, it
|
||||
is often clearer to pass the graph to be launched explicitly to
|
||||
the session constructor.
|
||||
|
||||
Args:
|
||||
target: Optional. The TensorFlow execution engine to connect to.
|
||||
graph: Optional. The `Graph` object to be used. If this argument is None,
|
||||
the default graph will be used.
|
||||
target: (Optional.) The execution engine to connect to.
|
||||
Defaults to using an in-process engine. At present, no value
|
||||
other than the empty string is supported.
|
||||
graph: (Optional.) The `Graph` to be launched (described above).
|
||||
"""
|
||||
super(InteractiveSession, self).__init__(target, graph)
|
||||
self._default_session = self.as_default()
|
||||
|
||||
@ -79,8 +79,7 @@ def all_libraries(module_to_name, members, documented):
|
||||
"all_candidate_sampler",
|
||||
"embedding_lookup_sparse"],
|
||||
prefix=PREFIX_TEXT),
|
||||
library('client', "Running Graphs", client_lib,
|
||||
exclude_symbols=["InteractiveSession"]),
|
||||
library('client', "Running Graphs", client_lib),
|
||||
library("train", "Training", tf.train,
|
||||
exclude_symbols=["Feature", "Features", "BytesList", "FloatList",
|
||||
"Int64List", "Example", "InferenceExample",
|
||||
|
||||
@ -56,17 +56,13 @@ RUN mkdir /bazel && \
|
||||
# Download and build TensorFlow.
|
||||
WORKDIR /tensorflow
|
||||
|
||||
# We can't clone the TF git repo yet, because of permissions issues.
|
||||
# RUN git clone https://tensorflow.googlesource.com/
|
||||
# Instead, we manually copy it in:
|
||||
# Add in the source tree
|
||||
COPY tensorflow /tensorflow
|
||||
|
||||
# Now we build
|
||||
RUN bazel clean && \
|
||||
bazel build -c opt tensorflow/tools/docker:simple_console
|
||||
bazel build -c opt tensorflow/tools/pip_package:build_pip_package && \
|
||||
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/pip && \
|
||||
pip install /tmp/pip/tensorflow-*.whl
|
||||
|
||||
ENV PYTHONPATH=/tensorflow/bazel-bin/tensorflow/tools/docker/simple_console.runfiles/:$PYTHONPATH
|
||||
|
||||
# We want to start Jupyter in the directory with our getting started
|
||||
# tutorials.
|
||||
WORKDIR /notebooks
|
||||
WORKDIR /root
|
||||
|
||||
Loading…
Reference in New Issue
Block a user