Need help with Optimus?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

225 Stars 24 Forks 85 Commits 8 Opened issues


Optimus: the first large-scale pre-trained VAE language model

Services available


Need anything else?

Contributors list

# 167,107
82 commits

Optimus: the first pre-trained Big VAE language model

This repository contains source code necessary to reproduce the results presented in the EMNLP 2020 paper Optimus: Organizing Sentences via Pre-trained Modeling of a Latent Space.

| | |-------------------------|:-------------------------:| | The network architecture of Optimus: encoder for representation learning and decoder for generation | Sentences are organized and manipulated in a pre-trained compact and smooth latent space

For more on this project, see the Microsoft Research Blog post.


May 21, 2020: Releasing a

for latent space manipulation, including sentence interpolation and analogy. Check out the

May 20, 2020: The latent space manipulation code is cleaned and released. See instructions at

May 13, 2020: The fine-tuning code for langauge modeling is released. See instructions at


There are four steps to use this codebase to reproduce the results in the paper.

  1. Dependencies
  2. Prepare datasets
  3. Model training
    1. Pre-training on setences in Wikipedia
    2. Languange Modeling
    3. Guided Language Generation
    4. Low-resource Language Understanding
  4. Collect and plot results


Pull docker from Docker Hub at:

. Please see the instruction at

The project is organized into the following structures, with ensential files & folders visualized.

saves the models checkpoints.
├── Optimus
   └── code
       ├── examples
           ├── big_ae
               ├── modules
                   └── ...
               └── ...
       ├── pytorch_transformers
               └── ...
       ├── scripts
           ├── scripts_docker
       ├── scripts_local
       ├── scripts_philly
   └── data
       └── datasets
           ├── wikipedia_json_64_filtered
               └── ...
       ├── snli_data
           └── ...
   └── output
       ├── pretrain
       ├── LM
       └── ...       

Prepare Datasets

Please download or preparation the data via following the instructions at


Model Training

1. Pre-training on setences in Wikipedia

We pre-trained our models on Philly (a Microsoft internal compute cluster), the code is specialized for multi-node multi-GPU compute on this platform. The pre-training main python is
. You may need to adjust the distributed training scripts.

2. Languange Modeling

To have a fair comparison with existing VAE languange models, we consider a model with latent dimension 32. The pre-trained model is fine-tuned on four commonly datasets for one epoch. Please see the details at


3. Guided Language Generation

Latent Space Manipulation To ensure good performance, we consider a model with latent dimension 768. The pre-trained model is fine-tuned on SNLI dataset, where sentences show related patterns. Please see the details at Please see the details at


4. Low-resource Language Understanding

Collect and Plot Results

Once the networks are trained and the results are saved, we extracted key results using Python script. The results can be plotted using the included IPython notebook

. Start the IPython Notebook server:
$ cd plots
$ ipython notebook

Select the

notebook and execute the included code. Note that without modification, we have copyed our extracted results into the notebook, and script will output figures in the paper. If you've run your own training and wish to plot results, you'll have to organize your results in the same format instead.


Please drop me (Chunyuan) a line if you have any questions.

  title={Optimus: Organizing Sentences via Pre-trained Modeling of a Latent Space},
  author={Li, Chunyuan and Gao, Xiang and Li, Yuan and Li, Xiujun and Peng, Baolin and Zhang, Yizhe and Gao, Jianfeng},

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.