Rooba
/
spaCy
mirror of https://github.com/explosion/spaCy.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity
9,823 Commits 68 Branches 175 Tags 298 MiB
Commit Graph

31 Commits

Author SHA1 Message Date
Matthew Honnibal 82277f63a3 💫 Small efficiency fixes to tokenizer (#2587) 
This patch improves tokenizer speed by about 10%, and reduces memory usage in the `Vocab` by removing a redundant index. The `vocab._by_orth` and `vocab._by_hash` indexed on different data in v1, but in v2 the orth and the hash are identical.

The patch also fixes an uninitialized variable in the tokenizer, the `has_special` flag. This checks whether a chunk we're tokenizing triggers a special-case rule. If it does, then we avoid caching within the chunk. This check led to incorrectly rejecting some chunks from the cache. 

With the `en_core_web_md` model, we now tokenize the IMDB train data at 503,104k words per second. Prior to this patch, we had 465,764k words per second.

Before switching to the regex library and supporting more languages, we had 1.3m words per second for the tokenizer. In order to recover the missing speed, we need to:

* Fix the variable-length lookarounds in the suffix, infix and `token_match` rules
* Improve the performance of the `token_match` regex
* Switch back from the `regex` library to the `re` library.

## Checklist
<!--- Before you submit the PR, go over this checklist and make sure you can
tick off all the boxes. [] -> [x] -->
- [x] I have submitted the spaCy Contributor Agreement.
- [x] I ran the tests, and all new and existing tests passed.
- [x] My changes don't require a change to the documentation, or if they do, I've added all required information.
 2018-07-24 23:35:54 +02:00
Explosion Bot 7b56b2f04b Add Vocab.cfg attr, to hold stuff like oov probs 2017-10-30 16:08:50 +01:00
Matthew Honnibal a131981f3b Work on vectors 2017-05-30 23:34:50 +02:00
Matthew Honnibal 9e167b7bb6 Strip serializer from code 2017-05-09 17:28:50 +02:00
Matthew Honnibal ca32a1ab01 Revert "Work on Issue #285: intern strings into document-specific pools, to address streaming data memory growth. StringStore.__getitem__ now raises KeyError when it can't find the string. Use StringStore.intern() to get the old behaviour. Still need to hunt down all uses of StringStore.__getitem__ in library and do testing, but logic looks good." 
This reverts commit 8423e8627f.
 2016-09-30 20:20:22 +02:00
Matthew Honnibal 8423e8627f Work on Issue #285: intern strings into document-specific pools, to address streaming data memory growth. StringStore.__getitem__ now raises KeyError when it can't find the string. Use StringStore.intern() to get the old behaviour. Still need to hunt down all uses of StringStore.__getitem__ in library and do testing, but logic looks good. 2016-09-30 10:14:47 +02:00
Matthew Honnibal 95aaea0d3f Refactor so that the tokenizer data is read from Python data, rather than from disk 2016-09-25 14:49:53 +02:00
Matthew Honnibal 85e7944572 * Start trying to pickle Vocab 2015-10-13 13:44:41 +11:00
Matthew Honnibal 362526b592 * Rename vectors_length attribute 2015-09-15 14:43:31 +10:00
Matthew Honnibal e285ca7d6c * Load serializer freqs in vocab 2015-09-10 15:22:48 +02:00
Matthew Honnibal 86c888667f * Merge in changes from de branch 2015-09-06 19:49:28 +02:00
Matthew Honnibal d2fc104a26 * Begin merge of Gazetteer and DE branches 2015-09-06 19:45:15 +02:00
Matthew Honnibal c2307fa9ee * More work on language-generic parsing 2015-08-28 02:02:33 +02:00
Matthew Honnibal 2d521768a3 * Store Morphology class in Vocab 2015-08-26 19:21:03 +02:00
Matthew Honnibal 6f1743692a * Work on language-independent refactoring 2015-08-23 20:49:18 +02:00
Matthew Honnibal fd525f0675 * Pass OOV probability around 2015-07-25 23:29:51 +02:00
Matthew Honnibal a7c4d72e83 * Add serializer property to Vocab, and lazy-load it. Add get_by_orth method. 2015-07-23 01:18:19 +02:00
Matthew Honnibal 109106a949 * Replace UniStr, using unicode objects instead 2015-07-22 04:52:05 +02:00
Matthew Honnibal 317cbbc015 * Serialization round trip now working with decent API, but with rough spots in the organisation and requiring vocabulary to be fixed ahead of time. 2015-07-19 15:18:17 +02:00
Matthew Honnibal 82d84b0f2b * Index lexemes by orth, instead of a lexemes vector. Breaks the mechanism for deciding not to own LexemeC structs during parsing. Need to reinstate this. 2015-07-18 22:42:15 +02:00
Matthew Honnibal 4dddc8a69b * Fix type declarations for attr_t. Remove unused id_t. 2015-07-18 22:39:57 +02:00
Matthew Honnibal db9dfd2e23 * Major refactor of serialization. Nearly complete now. 2015-07-17 01:27:54 +02:00
Matthew Honnibal af5cc926a4 * Add codec property to Vocab, to use the Huffman encoding 2015-07-13 13:55:14 +02:00
Matthew Honnibal abc43b852d * Add pos_tags attr to Vocab. 2015-07-08 12:36:38 +02:00
Matthew Honnibal c04e6ebca6 * Allow user to load different sized vectors. 2015-06-05 16:26:39 +02:00
Jordan Suchow 3a8d9b37a6 Remove trailing whitespace 2015-04-19 13:01:38 -07:00
Matthew Honnibal 0930892fc1 * Tmp. Working on refactor. Compiles, must hook up lexical feats. 2015-01-14 00:03:48 +11:00
Matthew Honnibal ce2edd6312 * Tmp commit. Refactoring to create a Python Lexeme class. 2015-01-12 10:26:22 +11:00
Matthew Honnibal b8b65903fc * Tmp 2014-12-24 17:42:00 +11:00
Matthew Honnibal d11c1edf8c * Import slice_unicode from strings.pyx 2014-12-20 07:56:26 +11:00
Matthew Honnibal 116f7f3bc1 * Rename Lexicon to Vocab, and move it to its own file 2014-12-20 06:54:03 +11:00