Skip to content

Commit

Permalink
chore(ROADMAP): mark word vector integration as completed
Browse files Browse the repository at this point in the history
  • Loading branch information
sanjayaksaxena committed Apr 26, 2024
1 parent fe891c6 commit aff4bb1
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion ROADMAP.md
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@ With winkNLP's production ready release in late 2020, the core is already in pla
|---|---|---|---|
|01.|**Extractive Summarization**:<br/> Add `its.sentenceWiseImprotance` helper to extract sentence wise impotance from a document. This may be used for extractive summarization apart from other usage. While it should be language agnostic, but it should leverage loaded language model's capability to improve summarization.| Simple | [Completed](https://observablehq.com/@winkjs/how-to-visualize-key-sentences-in-a-document) |
|02.|**Text Pre-processor**:<br/>Add a text preprocessing utility that provides options to (a) filter specific tokens based on their properties such as `pos`, `isStopWordFlag`, and `type`; (b) map entity type with a definable keyword; (c) add bigrams & trigrams and (d) inject sentiment. The API should follow winkNLP style and standards.|Medium|YTS|
|03.|**Word Vectors Integration**:<br/>Add integration with various word vectors starting with GloVe. This should include compression/decompression for fast loading, helpers for token, sentence and document vector computation. |High|WIP|
|03.|**Word Vectors Integration**:<br/>Add integration with various word vectors starting with GloVe. This should include compression/decompression for fast loading, helpers for token, sentence and document vector computation. |High|[Completed](https://github.com/winkjs/wink-nlp/releases/tag/2.0.0)|
|04.|**Sub-word Tokenizer**:<br/>Add sub-word tokenization feature using techniques like Byte Pair Encoding (BPE) and/or WordPiece. The processing pipeline should allow choice of tokenizer.|Very High|YTS|
|05.|**Compose Corpus**:<br/>Add a utility to produce training corpus using patterns and cartesian product.|Simple|YTS|
|06.|**Keywords Extraction**:<br/>Add `its.keywords` helper to extract keywords/keyphrases from the text via `doc.out( its.keywords )`. While it should be language agnostic, but it should leverage loaded language model's capability to improve extraction.| Simple | YTS |
Expand Down

0 comments on commit aff4bb1

Please sign in to comment.