https://github.com/sanyam5/skip-thoughts
The first public PyTorch implementation of Skip-Thought Vectors
https://github.com/sanyam5/skip-thoughts
Last synced: 25 days ago
JSON representation
The first public PyTorch implementation of Skip-Thought Vectors
- Host: GitHub
- URL: https://github.com/sanyam5/skip-thoughts
- Owner: sanyam5
- License: mit
- Created: 2017-12-17T06:17:30.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-01-19T07:18:08.000Z (over 7 years ago)
- Last Synced: 2024-11-14T05:34:29.669Z (7 months ago)
- Language: Python
- Homepage:
- Size: 27.3 KB
- Stars: 226
- Watchers: 7
- Forks: 47
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-pytorch-list-CNVersion - skip-thoughts - Thought词向量模型。 (Tutorials & books & examples|教程 & 书籍 & 示例 / Other libraries|其他库:)
- Awesome-pytorch-list - skip-thoughts - Thought Vectors in PyTorch. (Tutorials, books, & examples / Other libraries:)
README
# skip-thoughts
An implementation of Skip-Thought Vectors in PyTorch.### Blog
Here's a [blog explaining the subtleties of Skip-Thoughts](http://sanyam5.github.io/my-thoughts-on-skip-thoughts/)

Instructions
------------### Training
* Download [BookCorpus](http://yknzhu.wixsite.com/mbweb) or any other data-set and concatenate all sentences into one file and put it in `./data/` directory
* Modify the following line in `Train.ipynb` notebook accordingly:
`d = DataLoader("./data/dummy_corpus.txt")`
* There is no early stopping.
* The `Train` notebook runs at the rate of 1 epoch / 2 days on an Nvidia 1080 Ti.
* Your model is saved when `./saved_models` when the average training loss in the last 20 iterations dips below the previous best.### Evaluation
Only implemented on classification tasks
* Download the [movie review dataset](http://www.cs.cornell.edu/people/pabo/movie-review-data/rt-polaritydata.tar.gz) and put `rt-polarity.neg` and `rt-polarity.pos` in the `./tasks/mr_data` directory.
* You may also test on other classification tasks by downloading the datasets and providing their path and tasks type in `Evaluate.ipynb`