fasttext-embeddings-with-flair
To get a better overview and understanding of the need for this project, I would recommend to read my article on the same.
The objective of this script is to be able to use our custom fasttext embeddings with the Flair.
The integration is very simple:
Copy the fasttext_custom_embeddings_with_flair.py
file to your project.
Install the packages from requirements.txt
.
Import the FastTextEmbeddings
from fasttext_custom_embeddings_with_flair.py
file.
Instantiate an object of FastTextEmbeddings
by passing either the local path or the remote http(s) url to the constructor.
Use it like you use any other embedding object in Flair.
from fasttext_custom_embeddings_with_flair import FastTextEmbeddings from flair.data import Sentence ft_embeddings = FastTextEmbeddings('/path/to/custom_fasttext_embeddings.bin', use_local=True) sentence = Sentence('The quick brown fox jumps over a lazy dog.', use_tokenizer=True) 'ft_embeddings.embed(sentence)
Initialize the constructor by passing the local path to embeddings .bin
file.
Set the boolean use_local
to True
.
Support for embedding file with ".vec" extension is also added. We need to add one more parameter extension='.vec'
to the constructor.
ft_embeddings = FastTextEmbeddings('/path/to/custom_fasttext_embeddings.vec', use_local=True, extension='.vec')
Note: The ".vec" extension files do not support out of vocabulary(OOV) embeddings. In this case, for OOV words, zero vectors are assigned by default.
from fasttext_custom_embeddings_with_flair import FastTextEmbeddings from flair.data import Sentence ft_embeddings = FastTextEmbeddings('/url/to/custom_fattext_embeddings.bin', use_local=False) sentence = Sentence('The quick brown fox jumps over a lazy dog.', use_tokenizer=True) ft_embeddings.embed(sentence)
Initialize the constructor by passing the remote url to embeddings .bin
file. (I have tested this with my embeddings in S3 with an https url.)
Set the boolean use_local
to False
. (Mandatory step if you are using a remote url.)
from fasttext_custom_embeddings_with_flair import FastTextEmbeddings from flair.embeddings import WordEmbeddings, StackedEmbeddingsfrom flair.data import Sentence ft_embeddings = FastTextEmbeddings('/url/to/custom_fattext_embeddings.bin', use_local=False) glove_embeddings = WordEmbeddings('glove') stacked_embeddings = StackedEmbeddings([ft_embeddings, glove_embeddings]) sentence = Sentence('The quick brown fox jumps over a lazy dog.', use_tokenizer=True) stacked_embeddings.embed(sentence)
from fasttext_custom_embeddings_with_flair import FastTextEmbeddings from flair.embeddings import WordEmbeddings, DocumentRNNEmbeddings from flair.data import Sentence ft_embeddings = FastTextEmbeddings('/url/to/custom_fattext_embeddings.bin', use_local=True) glove_embeddings = WordEmbeddings('glove') document_rnn_embeddings = DocumentRNNEmbeddings([ft_embeddings, glove_embeddings]) sentence = Sentence('The quick brown fox jumps over a lazy dog.', use_tokenizer=True) document_rnn_embeddings.embed(sentence)
In case of any doubts, get in touch.
Author: Pranay Chandekar
Refernces:
This repository is licensed under the Apache 2.0 License.
下一篇:Flair-JSON-NLP
还没有评论,说两句吧!
热门资源
Keras-ResNeXt
Keras ResNeXt Implementation of ResNeXt models...
seetafaceJNI
项目介绍 基于中科院seetaface2进行封装的JAVA...
spark-corenlp
This package wraps Stanford CoreNLP annotators ...
capsnet-with-caps...
CapsNet with capsule-wise convolution Project ...
inferno-boilerplate
This is a very basic boilerplate example for pe...
智能在线
400-630-6780
聆听.建议反馈
E-mail: support@tusaishared.com