fasttext-wheel-woa 0.9.2__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (52) hide show
  1. fasttext_wheel_woa-0.9.2/LICENSE +21 -0
  2. fasttext_wheel_woa-0.9.2/MANIFEST.in +5 -0
  3. fasttext_wheel_woa-0.9.2/PKG-INFO +442 -0
  4. fasttext_wheel_woa-0.9.2/README.md +339 -0
  5. fasttext_wheel_woa-0.9.2/pyproject.toml +2 -0
  6. fasttext_wheel_woa-0.9.2/python/README.md +322 -0
  7. fasttext_wheel_woa-0.9.2/python/README.rst +406 -0
  8. fasttext_wheel_woa-0.9.2/python/benchmarks/README.rst +3 -0
  9. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/FastText.py +573 -0
  10. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/__init__.py +22 -0
  11. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/pybind/fasttext_pybind.cc +526 -0
  12. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/tests/__init__.py +14 -0
  13. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/tests/test_configurations.py +239 -0
  14. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/tests/test_script.py +629 -0
  15. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/util/__init__.py +15 -0
  16. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext/util/util.py +209 -0
  17. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/PKG-INFO +442 -0
  18. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/SOURCES.txt +58 -0
  19. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/dependency_links.txt +1 -0
  20. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/not-zip-safe +1 -0
  21. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/requires.txt +3 -0
  22. fasttext_wheel_woa-0.9.2/python/fasttext_module/fasttext_wheel_woa.egg-info/top_level.txt +2 -0
  23. fasttext_wheel_woa-0.9.2/setup.cfg +7 -0
  24. fasttext_wheel_woa-0.9.2/setup.py +203 -0
  25. fasttext_wheel_woa-0.9.2/src/args.cc +493 -0
  26. fasttext_wheel_woa-0.9.2/src/args.h +96 -0
  27. fasttext_wheel_woa-0.9.2/src/autotune.cc +477 -0
  28. fasttext_wheel_woa-0.9.2/src/autotune.h +89 -0
  29. fasttext_wheel_woa-0.9.2/src/densematrix.cc +175 -0
  30. fasttext_wheel_woa-0.9.2/src/densematrix.h +83 -0
  31. fasttext_wheel_woa-0.9.2/src/dictionary.cc +540 -0
  32. fasttext_wheel_woa-0.9.2/src/dictionary.h +111 -0
  33. fasttext_wheel_woa-0.9.2/src/fasttext.cc +832 -0
  34. fasttext_wheel_woa-0.9.2/src/fasttext.h +170 -0
  35. fasttext_wheel_woa-0.9.2/src/loss.cc +346 -0
  36. fasttext_wheel_woa-0.9.2/src/loss.h +163 -0
  37. fasttext_wheel_woa-0.9.2/src/main.cc +454 -0
  38. fasttext_wheel_woa-0.9.2/src/matrix.cc +25 -0
  39. fasttext_wheel_woa-0.9.2/src/matrix.h +44 -0
  40. fasttext_wheel_woa-0.9.2/src/meter.cc +214 -0
  41. fasttext_wheel_woa-0.9.2/src/meter.h +91 -0
  42. fasttext_wheel_woa-0.9.2/src/model.cc +97 -0
  43. fasttext_wheel_woa-0.9.2/src/model.h +79 -0
  44. fasttext_wheel_woa-0.9.2/src/productquantizer.cc +251 -0
  45. fasttext_wheel_woa-0.9.2/src/productquantizer.h +63 -0
  46. fasttext_wheel_woa-0.9.2/src/quantmatrix.cc +117 -0
  47. fasttext_wheel_woa-0.9.2/src/quantmatrix.h +60 -0
  48. fasttext_wheel_woa-0.9.2/src/real.h +14 -0
  49. fasttext_wheel_woa-0.9.2/src/utils.cc +53 -0
  50. fasttext_wheel_woa-0.9.2/src/utils.h +72 -0
  51. fasttext_wheel_woa-0.9.2/src/vector.cc +96 -0
  52. fasttext_wheel_woa-0.9.2/src/vector.h +61 -0
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2016-present, Facebook, Inc.
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
@@ -0,0 +1,5 @@
1
+ include LICENSE
2
+ include PATENTS
3
+
4
+ recursive-include python *.md *.rst
5
+ recursive-include src *.h
@@ -0,0 +1,442 @@
1
+ Metadata-Version: 2.4
2
+ Name: fasttext-wheel-woa
3
+ Version: 0.9.2
4
+ Summary: fasttext Python bindings - Windows on ARM (ARM64) fork with native build fixes
5
+ Home-page: https://github.com/facebookresearch/fastText
6
+ Author: kirinekoneko
7
+ Author-email: kirineko@qq.com
8
+ License: MIT
9
+ Classifier: Development Status :: 3 - Alpha
10
+ Classifier: Intended Audience :: Developers
11
+ Classifier: Intended Audience :: Science/Research
12
+ Classifier: License :: OSI Approved :: MIT License
13
+ Classifier: Programming Language :: Python :: 2.7
14
+ Classifier: Programming Language :: Python :: 3.4
15
+ Classifier: Programming Language :: Python :: 3.5
16
+ Classifier: Programming Language :: Python :: 3.6
17
+ Classifier: Topic :: Software Development
18
+ Classifier: Topic :: Scientific/Engineering
19
+ Classifier: Operating System :: Microsoft :: Windows
20
+ Classifier: Operating System :: POSIX
21
+ Classifier: Operating System :: Unix
22
+ Classifier: Operating System :: MacOS
23
+ License-File: LICENSE
24
+ Requires-Dist: pybind11>=2.2
25
+ Requires-Dist: setuptools>=0.7.0
26
+ Requires-Dist: numpy
27
+ Dynamic: author
28
+ Dynamic: author-email
29
+ Dynamic: classifier
30
+ Dynamic: description
31
+ Dynamic: home-page
32
+ Dynamic: license
33
+ Dynamic: license-file
34
+ Dynamic: requires-dist
35
+ Dynamic: summary
36
+
37
+ fastText |CircleCI|
38
+ ===================
39
+
40
+ `fastText <https://fasttext.cc/>`__ is a library for efficient learning
41
+ of word representations and sentence classification.
42
+
43
+ In this document we present how to use fastText in python.
44
+
45
+ Table of contents
46
+ -----------------
47
+
48
+ - `Requirements <#requirements>`__
49
+ - `Installation <#installation>`__
50
+ - `Usage overview <#usage-overview>`__
51
+ - `Word representation model <#word-representation-model>`__
52
+ - `Text classification model <#text-classification-model>`__
53
+ - `IMPORTANT: Preprocessing data / encoding
54
+ conventions <#important-preprocessing-data-encoding-conventions>`__
55
+ - `More examples <#more-examples>`__
56
+ - `API <#api>`__
57
+ - `train_unsupervised parameters <#train_unsupervised-parameters>`__
58
+ - `train_supervised parameters <#train_supervised-parameters>`__
59
+ - `model object <#model-object>`__
60
+
61
+ Requirements
62
+ ============
63
+
64
+ `fastText <https://fasttext.cc/>`__ builds on modern Mac OS and Linux
65
+ distributions. Since it uses C++11 features, it requires a compiler with
66
+ good C++11 support. You will need `Python <https://www.python.org/>`__
67
+ (version 2.7 or ≥ 3.4), `NumPy <http://www.numpy.org/>`__ &
68
+ `SciPy <https://www.scipy.org/>`__ and
69
+ `pybind11 <https://github.com/pybind/pybind11>`__.
70
+
71
+ Installation
72
+ ============
73
+
74
+ To install the latest release, you can do :
75
+
76
+ .. code:: bash
77
+
78
+ $ pip install fasttext
79
+
80
+ or, to get the latest development version of fasttext, you can install
81
+ from our github repository :
82
+
83
+ .. code:: bash
84
+
85
+ $ git clone https://github.com/facebookresearch/fastText.git
86
+ $ cd fastText
87
+ $ sudo pip install .
88
+ $ # or :
89
+ $ sudo python setup.py install
90
+
91
+ Usage overview
92
+ ==============
93
+
94
+ Word representation model
95
+ -------------------------
96
+
97
+ In order to learn word vectors, as `described
98
+ here <https://fasttext.cc/docs/en/references.html#enriching-word-vectors-with-subword-information>`__,
99
+ we can use ``fasttext.train_unsupervised`` function like this:
100
+
101
+ .. code:: py
102
+
103
+ import fasttext
104
+
105
+ # Skipgram model :
106
+ model = fasttext.train_unsupervised('data.txt', model='skipgram')
107
+
108
+ # or, cbow model :
109
+ model = fasttext.train_unsupervised('data.txt', model='cbow')
110
+
111
+ where ``data.txt`` is a training file containing utf-8 encoded text.
112
+
113
+ The returned ``model`` object represents your learned model, and you can
114
+ use it to retrieve information.
115
+
116
+ .. code:: py
117
+
118
+ print(model.words) # list of words in dictionary
119
+ print(model['king']) # get the vector of the word 'king'
120
+
121
+ Saving and loading a model object
122
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
123
+
124
+ You can save your trained model object by calling the function
125
+ ``save_model``.
126
+
127
+ .. code:: py
128
+
129
+ model.save_model("model_filename.bin")
130
+
131
+ and retrieve it later thanks to the function ``load_model`` :
132
+
133
+ .. code:: py
134
+
135
+ model = fasttext.load_model("model_filename.bin")
136
+
137
+ For more information about word representation usage of fasttext, you
138
+ can refer to our `word representations
139
+ tutorial <https://fasttext.cc/docs/en/unsupervised-tutorial.html>`__.
140
+
141
+ Text classification model
142
+ -------------------------
143
+
144
+ In order to train a text classifier using the method `described
145
+ here <https://fasttext.cc/docs/en/references.html#bag-of-tricks-for-efficient-text-classification>`__,
146
+ we can use ``fasttext.train_supervised`` function like this:
147
+
148
+ .. code:: py
149
+
150
+ import fasttext
151
+
152
+ model = fasttext.train_supervised('data.train.txt')
153
+
154
+ where ``data.train.txt`` is a text file containing a training sentence
155
+ per line along with the labels. By default, we assume that labels are
156
+ words that are prefixed by the string ``__label__``
157
+
158
+ Once the model is trained, we can retrieve the list of words and labels:
159
+
160
+ .. code:: py
161
+
162
+ print(model.words)
163
+ print(model.labels)
164
+
165
+ To evaluate our model by computing the precision at 1 (P@1) and the
166
+ recall on a test set, we use the ``test`` function:
167
+
168
+ .. code:: py
169
+
170
+ def print_results(N, p, r):
171
+ print("N\t" + str(N))
172
+ print("P@{}\t{:.3f}".format(1, p))
173
+ print("R@{}\t{:.3f}".format(1, r))
174
+
175
+ print_results(*model.test('test.txt'))
176
+
177
+ We can also predict labels for a specific text :
178
+
179
+ .. code:: py
180
+
181
+ model.predict("Which baking dish is best to bake a banana bread ?")
182
+
183
+ By default, ``predict`` returns only one label : the one with the
184
+ highest probability. You can also predict more than one label by
185
+ specifying the parameter ``k``:
186
+
187
+ .. code:: py
188
+
189
+ model.predict("Which baking dish is best to bake a banana bread ?", k=3)
190
+
191
+ If you want to predict more than one sentence you can pass an array of
192
+ strings :
193
+
194
+ .. code:: py
195
+
196
+ model.predict(["Which baking dish is best to bake a banana bread ?", "Why not put knives in the dishwasher?"], k=3)
197
+
198
+ Of course, you can also save and load a model to/from a file as `in the
199
+ word representation usage <#saving-and-loading-a-model-object>`__.
200
+
201
+ For more information about text classification usage of fasttext, you
202
+ can refer to our `text classification
203
+ tutorial <https://fasttext.cc/docs/en/supervised-tutorial.html>`__.
204
+
205
+ Compress model files with quantization
206
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
207
+
208
+ When you want to save a supervised model file, fastText can compress it
209
+ in order to have a much smaller model file by sacrificing only a little
210
+ bit performance.
211
+
212
+ .. code:: py
213
+
214
+ # with the previously trained `model` object, call :
215
+ model.quantize(input='data.train.txt', retrain=True)
216
+
217
+ # then display results and save the new model :
218
+ print_results(*model.test(valid_data))
219
+ model.save_model("model_filename.ftz")
220
+
221
+ ``model_filename.ftz`` will have a much smaller size than
222
+ ``model_filename.bin``.
223
+
224
+ For further reading on quantization, you can refer to `this paragraph
225
+ from our blog
226
+ post <https://fasttext.cc/blog/2017/10/02/blog-post.html#model-compression>`__.
227
+
228
+ IMPORTANT: Preprocessing data / encoding conventions
229
+ ----------------------------------------------------
230
+
231
+ In general it is important to properly preprocess your data. In
232
+ particular our example scripts in the `root
233
+ folder <https://github.com/facebookresearch/fastText>`__ do this.
234
+
235
+ fastText assumes UTF-8 encoded text. All text must be `unicode for
236
+ Python2 <https://docs.python.org/2/library/functions.html#unicode>`__
237
+ and `str for
238
+ Python3 <https://docs.python.org/3.5/library/stdtypes.html#textseq>`__.
239
+ The passed text will be `encoded as UTF-8 by
240
+ pybind11 <https://pybind11.readthedocs.io/en/master/advanced/cast/strings.html?highlight=utf-8#strings-bytes-and-unicode-conversions>`__
241
+ before passed to the fastText C++ library. This means it is important to
242
+ use UTF-8 encoded text when building a model. On Unix-like systems you
243
+ can convert text using `iconv <https://en.wikipedia.org/wiki/Iconv>`__.
244
+
245
+ fastText will tokenize (split text into pieces) based on the following
246
+ ASCII characters (bytes). In particular, it is not aware of UTF-8
247
+ whitespace. We advice the user to convert UTF-8 whitespace / word
248
+ boundaries into one of the following symbols as appropiate.
249
+
250
+ - space
251
+ - tab
252
+ - vertical tab
253
+ - carriage return
254
+ - formfeed
255
+ - the null character
256
+
257
+ The newline character is used to delimit lines of text. In particular,
258
+ the EOS token is appended to a line of text if a newline character is
259
+ encountered. The only exception is if the number of tokens exceeds the
260
+ MAX\_LINE\_SIZE constant as defined in the `Dictionary
261
+ header <https://github.com/facebookresearch/fastText/blob/master/src/dictionary.h>`__.
262
+ This means if you have text that is not separate by newlines, such as
263
+ the `fil9 dataset <http://mattmahoney.net/dc/textdata>`__, it will be
264
+ broken into chunks with MAX\_LINE\_SIZE of tokens and the EOS token is
265
+ not appended.
266
+
267
+ The length of a token is the number of UTF-8 characters by considering
268
+ the `leading two bits of a
269
+ byte <https://en.wikipedia.org/wiki/UTF-8#Description>`__ to identify
270
+ `subsequent bytes of a multi-byte
271
+ sequence <https://github.com/facebookresearch/fastText/blob/master/src/dictionary.cc>`__.
272
+ Knowing this is especially important when choosing the minimum and
273
+ maximum length of subwords. Further, the EOS token (as specified in the
274
+ `Dictionary
275
+ header <https://github.com/facebookresearch/fastText/blob/master/src/dictionary.h>`__)
276
+ is considered a character and will not be broken into subwords.
277
+
278
+ More examples
279
+ -------------
280
+
281
+ In order to have a better knowledge of fastText models, please consider
282
+ the main
283
+ `README <https://github.com/facebookresearch/fastText/blob/master/README.md>`__
284
+ and in particular `the tutorials on our
285
+ website <https://fasttext.cc/docs/en/supervised-tutorial.html>`__.
286
+
287
+ You can find further python examples in `the doc
288
+ folder <https://github.com/facebookresearch/fastText/tree/master/python/doc/examples>`__.
289
+
290
+ As with any package you can get help on any Python function using the
291
+ help function.
292
+
293
+ For example
294
+
295
+ ::
296
+
297
+ +>>> import fasttext
298
+ +>>> help(fasttext.FastText)
299
+
300
+ Help on module fasttext.FastText in fasttext:
301
+
302
+ NAME
303
+ fasttext.FastText
304
+
305
+ DESCRIPTION
306
+ # Copyright (c) 2017-present, Facebook, Inc.
307
+ # All rights reserved.
308
+ #
309
+ # This source code is licensed under the MIT license found in the
310
+ # LICENSE file in the root directory of this source tree.
311
+
312
+ FUNCTIONS
313
+ load_model(path)
314
+ Load a model given a filepath and return a model object.
315
+
316
+ tokenize(text)
317
+ Given a string of text, tokenize it and return a list of tokens
318
+ [...]
319
+
320
+ API
321
+ ===
322
+
323
+ ``train_unsupervised`` parameters
324
+ ---------------------------------
325
+
326
+ .. code:: python
327
+
328
+ input # training file path (required)
329
+ model # unsupervised fasttext model {cbow, skipgram} [skipgram]
330
+ lr # learning rate [0.05]
331
+ dim # size of word vectors [100]
332
+ ws # size of the context window [5]
333
+ epoch # number of epochs [5]
334
+ minCount # minimal number of word occurences [5]
335
+ minn # min length of char ngram [3]
336
+ maxn # max length of char ngram [6]
337
+ neg # number of negatives sampled [5]
338
+ wordNgrams # max length of word ngram [1]
339
+ loss # loss function {ns, hs, softmax, ova} [ns]
340
+ bucket # number of buckets [2000000]
341
+ thread # number of threads [number of cpus]
342
+ lrUpdateRate # change the rate of updates for the learning rate [100]
343
+ t # sampling threshold [0.0001]
344
+ verbose # verbose [2]
345
+
346
+ ``train_supervised`` parameters
347
+ -------------------------------
348
+
349
+ .. code:: python
350
+
351
+ input # training file path (required)
352
+ lr # learning rate [0.1]
353
+ dim # size of word vectors [100]
354
+ ws # size of the context window [5]
355
+ epoch # number of epochs [5]
356
+ minCount # minimal number of word occurences [1]
357
+ minCountLabel # minimal number of label occurences [1]
358
+ minn # min length of char ngram [0]
359
+ maxn # max length of char ngram [0]
360
+ neg # number of negatives sampled [5]
361
+ wordNgrams # max length of word ngram [1]
362
+ loss # loss function {ns, hs, softmax, ova} [softmax]
363
+ bucket # number of buckets [2000000]
364
+ thread # number of threads [number of cpus]
365
+ lrUpdateRate # change the rate of updates for the learning rate [100]
366
+ t # sampling threshold [0.0001]
367
+ label # label prefix ['__label__']
368
+ verbose # verbose [2]
369
+ pretrainedVectors # pretrained word vectors (.vec file) for supervised learning []
370
+
371
+ ``model`` object
372
+ ----------------
373
+
374
+ ``train_supervised``, ``train_unsupervised`` and ``load_model``
375
+ functions return an instance of ``_FastText`` class, that we generaly
376
+ name ``model`` object.
377
+
378
+ This object exposes those training arguments as properties : ``lr``,
379
+ ``dim``, ``ws``, ``epoch``, ``minCount``, ``minCountLabel``, ``minn``,
380
+ ``maxn``, ``neg``, ``wordNgrams``, ``loss``, ``bucket``, ``thread``,
381
+ ``lrUpdateRate``, ``t``, ``label``, ``verbose``, ``pretrainedVectors``.
382
+ So ``model.wordNgrams`` will give you the max length of word ngram used
383
+ for training this model.
384
+
385
+ In addition, the object exposes several functions :
386
+
387
+ .. code:: python
388
+
389
+ get_dimension # Get the dimension (size) of a lookup vector (hidden layer).
390
+ # This is equivalent to `dim` property.
391
+ get_input_vector # Given an index, get the corresponding vector of the Input Matrix.
392
+ get_input_matrix # Get a copy of the full input matrix of a Model.
393
+ get_labels # Get the entire list of labels of the dictionary
394
+ # This is equivalent to `labels` property.
395
+ get_line # Split a line of text into words and labels.
396
+ get_output_matrix # Get a copy of the full output matrix of a Model.
397
+ get_sentence_vector # Given a string, get a single vector represenation. This function
398
+ # assumes to be given a single line of text. We split words on
399
+ # whitespace (space, newline, tab, vertical tab) and the control
400
+ # characters carriage return, formfeed and the null character.
401
+ get_subword_id # Given a subword, return the index (within input matrix) it hashes to.
402
+ get_subwords # Given a word, get the subwords and their indicies.
403
+ get_word_id # Given a word, get the word id within the dictionary.
404
+ get_word_vector # Get the vector representation of word.
405
+ get_words # Get the entire list of words of the dictionary
406
+ # This is equivalent to `words` property.
407
+ is_quantized # whether the model has been quantized
408
+ predict # Given a string, get a list of labels and a list of corresponding probabilities.
409
+ quantize # Quantize the model reducing the size of the model and it's memory footprint.
410
+ save_model # Save the model to the given path
411
+ test # Evaluate supervised model using file given by path
412
+ test_label # Return the precision and recall score for each label.
413
+
414
+ The properties ``words``, ``labels`` return the words and labels from
415
+ the dictionary :
416
+
417
+ .. code:: py
418
+
419
+ model.words # equivalent to model.get_words()
420
+ model.labels # equivalent to model.get_labels()
421
+
422
+ The object overrides ``__getitem__`` and ``__contains__`` functions in
423
+ order to return the representation of a word and to check if a word is
424
+ in the vocabulary.
425
+
426
+ .. code:: py
427
+
428
+ model['king'] # equivalent to model.get_word_vector('king')
429
+ 'king' in model # equivalent to `'king' in model.get_words()`
430
+
431
+ Join the fastText community
432
+ ---------------------------
433
+
434
+ - `Facebook page <https://www.facebook.com/groups/1174547215919768>`__
435
+ - `Stack
436
+ overflow <https://stackoverflow.com/questions/tagged/fasttext>`__
437
+ - `Google
438
+ group <https://groups.google.com/forum/#!forum/fasttext-library>`__
439
+ - `GitHub <https://github.com/facebookresearch/fastText>`__
440
+
441
+ .. |CircleCI| image:: https://circleci.com/gh/facebookresearch/fastText/tree/master.svg?style=svg
442
+ :target: https://circleci.com/gh/facebookresearch/fastText/tree/master