skipthoughts 0.0.1

Skipthoughts pretrained models for Pytorch

Skip-Thoughts.torch for Pytorcb

Skip-Thoughts.torch is a lightweight porting of skip-thought pretrained models from Theano to Pytorch.

Installation

1. python3 with anaconda
2. pytorch with/out CUDA

Install from pip

3. pip install skipthoughts

Install from repo

3. git clone https://github.com/Cadene/skip-thoughts.torch.git
4. cd skip-thoughts.torch/pytorch
5. python setup.py install

Available pretrained models

UniSkip

It uses the nn.GRU layer from torch with the cudnn backend. It is the fastest implementation, but the dropout is sampled after each time-step in the cudnn implementation... (equals bad regularization)

DropUniSkip

It uses the nn.GRUCell layer from torch with the cudnn backend. It is slightly slower than UniSkip, however the dropout is sampled once for all time-steps in a sequence (good regularization).

BayesianUniSkip

It uses a custom GRU layer with a torch backend. It is at least two times slower than UniSkip, however the dropout is sampled once for all time-steps for each Linear (best regularization).

BiSkip

Equivalent to UniSkip, but with a bi-sequential GRU.

Quick example

import torch
from torch.autograd import Variable
import sys
sys.path.append('skip-thoughts.torch/pytorch')
from skipthoughts import UniSkip

dir_st = 'data/skip-thoughts'
vocab = ['robots', 'are', 'very', 'cool', '<eos>', 'BiDiBu']
uniskip = UniSkip(dir_st, vocab)

input = Variable(torch.LongTensor([
    [1,2,3,4,0], # robots are very cool 0
    [6,2,3,4,5]  # bidibu are very cool <eos>
])) # <eos> token is optional
print(input.size()) # batch_size x seq_len

output_seq2vec = uniskip(input, lengths=[4,5])
print(output_seq2vec.size()) # batch_size x 2400

output_seq2seq = uniskip(input)
print(output_seq2seq.size()) # batch_size x seq_len x 2400