From 2420d944cbb71bbe4b3819b5ce351d1372f92b8f Mon Sep 17 00:00:00 2001 From: Matthew Honnibal Date: Tue, 4 Nov 2014 17:01:54 +1100 Subject: [PATCH] * Upd sales copy --- docs/source/index.rst | 36 ++++++++++++++++++------------------ 1 file changed, 18 insertions(+), 18 deletions(-) diff --git a/docs/source/index.rst b/docs/source/index.rst index e0339f1e2..97681bfd8 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -6,20 +6,19 @@ spaCy NLP Tokenizer and Lexicon ================================ -spaCy is a library for industrial strength NLP in Python and Cython. Its core -values are efficiency, accuracy and minimalism. +spaCy is a library for industrial strength NLP in Python. Its core +values are: -* Efficiency: spaCy is TODOx faster than the Stanford tools, and TODOx faster - than NLTK. You won't find faster NLP tools. Using spaCy will save you - thousands in server costs, and will force you to make fewer compromises. +* **Efficiency**: You won't find faster NLP tools. For shallow analysis, it's 10x + faster than Stanford Core NLP, and over 200x faster than NLTK. Its parser is + over 100x faster than Stanford's. -* Accuracy: All spaCy tools are within 0.5% of the current published +* **Accuracy**: All spaCy tools are within 0.5% of the current published state-of-the-art, on both news and web text. NLP moves fast, so always check the numbers --- and don't settle for tools that aren't backed by - rigorous recent evaluation. An algorithm that was "close enough to state-of-the-art" - 5 years ago is probably crap by today's standards. + rigorous recent evaluation. -* Minimalism: This isn't a library that covers 43 known algorithms to do X. You +* **Minimalism**: This isn't a library that covers 43 known algorithms to do X. You get 1 --- the best one --- with a simple, low-level interface. This keeps the code-base small and concrete. Our Python APIs use lists and dictionaries, and our C/Cython APIs use arrays and simple structs. @@ -27,15 +26,16 @@ values are efficiency, accuracy and minimalism. Comparison ---------- -+-------------+-------------+---+-----------+--------------+ -| POS taggers | Speed (w/s) | % Acc. (news) | % Acc. (web) | -+-------------+-------------+---------------+--------------+ -| spaCy | | | | -+-------------+-------------+---------------+--------------+ -| Stanford | 16,000 | | | -+-------------+-------------+---------------+--------------+ -| NLTK | | | | -+-------------+-------------+---------------+--------------+ + ++----------------+-------------+--------+---------------+--------------+ +| Tokenize & Tag | Speed (w/s) | Memory | % Acc. (news) | % Acc. (web) | ++----------------+-------------+--------+---------------+--------------+ +| spaCy | 107,000 | 1.3gb | 96.7 | | ++----------------+-------------+--------+---------------+--------------+ +| Stanford | 8,000 | 1.5gb | 96.7 | | ++----------------+-------------+--------+---------------+--------------+ +| NLTK | 543 | 61mb | 94.0 | | ++----------------+-------------+--------+---------------+--------------+ .. toctree::