cotk 0.0.1

Conversational Toolkits

Conversational Toolkits

cotk is an open-source lightweight framework for model building and evaluation. We provides standard dataset and evaluation suites in the domain of general language generation. It easy to use and make you focus on designing your models!

Features included:

• Light-weight, easy to start. Don't bother your way to construct models.
• Predefined standard datasets, in the domain of language modeling, dialog generation and more.
• Predefined evaluation suites, test your model with multiple metrics in several lines.
• A dashboard to show experiments, compare your and others' models fairly.
• Long-term maintenance and consistent development.

This project is a part of dialtk (Toolkits for Dialog System by Tsinghua University), you can follow dialtk or cotk on our home page.

Quick links

• Tutorial & Documents
• Dashboard

Index

• Installation
  • Requirements
  • Install from pip
  • Install from source
• Quick Start
  • Dataloader
  • Metrics
  • Publish Experiments
  • Reproduce Experiments
  • Predefined Models
• Issues
• Contributions
• Team
• License

Installation

Requirements

• python 3
• numpy >= 1.13
• nltk >= 3.4
• tqdm >= 4.30
• checksumdir >= 1.1
• pytorch >= 1.0.0 (optional)
• pytorch-pretrained-bert (optional)

Install from pip

You can simply get the latest stable version from pip using

    pip install cotk

Install from source code

• Clone the cotk repository

    git clone https://github.com/thu-coai/cotk.git

• Install cotk via pip

    cd cotk
    pip install -e .

• If you want to run the models in ./models, you have to additionally install TensorFlow or PyTorch.

Quick Start

Let us skim through the whole package to find what you want.

Dataloader

Load common used dataset and preprocess for you:

• Download online resources or import from local
• Split training set, development set and test set
• Construct vocabulary list

    >>> # automatically download online resources
    >>> dataloader = cotk.dataloader.MSCOCO("resources://MSCOCO_small")
    >>> # or download from a url
    >>> dl_url = cotk.dataloader.MSCOCO("http://cotk-data.s3-ap-northeast-1.amazonaws.com/mscoco_small.zip#MSCOCO")
    >>> # or import from local file
    >>> dl_zip = cotk.dataloader.MSCOCO("./MSCOCO.zip#MSCOCO")

    >>> print("Dataset is split into:", dataloader.key_name)
    ["train", "dev", "test"]

Inspect vocabulary list

    >>> print("Vocabulary size:", dataloader.vocab_size)
    Vocabulary size: 2588
    >>> print("Frist 10 tokens in vocabulary:", dataloader.vocab_list[:10])
    Frist 10 tokens in vocabulary: ['<pad>', '<unk>', '<go>', '<eos>', '.', 'a', 'A', 'on', 'of', 'in']

Convert between ids and strings

    >>> print("Convert string to ids", \
    ...           dataloader.convert_tokens_to_ids(["<go>", "hello", "world", "<eos>"]))
    Convert string to string [2, 1379, 1897, 3]
    >>> print("Convert ids to string", \
    ...           dataloader.convert_ids_to_tokens([2, 1379, 1897, 3]))

Iterate over batches

    >>> for data in dataloader.get_batch("train", batch_size=1):
    ...     print(data)
    {'sent':
        array([[ 2, 181, 13, 26, 145, 177, 8, 22, 12, 5, 1, 1099, 4, 3]]),
        # <go> This is an old photo of people and a <unk> wagon.
     'sent_allvocabs':
        array([[ 2, 181, 13, 26, 145, 177, 8, 22, 12, 5, 3755, 1099, 4, 3]]),
        # <go> This is an old photo of people and a horse-drawn wagon.
     'sent_length': array([14])}
    ......

or using while if you like

    >>> dataloader.restart("train", batch_size=1):
    >>> while True:
    ...    data = dataloader.get_next_batch("train")
    ...    if data is None: break
    ...    print(data)

note: If you want to know more about data loader, please refer to docs.

Metrics

We found there are different versions of the same metric in released codes on Github, which leads to unfair compare between models. For example, whether considering unk, calculating the mean of NLL across sentences or tokens in perplexity may introduce an error of several times and extremely harm the evaluation.

We provide unified metrics implementation for all models. The metric object receives data in batch.

    >>> metric = cotk.metric.SelfBleuCorpusMetric(dataloader, gen_key="gen")
    >>> metric.forward({
    ...    "gen":
    ...        [[2, 181, 13, 26, 145, 177, 8, 22, 12, 5, 3755, 1099, 4, 3],
    ...         [2, 46, 145, 500, 1764, 207, 11, 5, 93, 7, 31, 4, 3]]
    ... })
    >>> print(metric.close())
    {'self-bleu': 0.02206768072402293,
     'self-bleu hashvalue': 'c206893c2272af489147b80df306ee703e71d9eb178f6bb06c73cb935f474452'}

We also provide standard metrics for selected dataloader.

    >>> metric = dataloader.get_inference_metric(gen_key="gen")
    >>> metric.forward({
    ...    "gen":
    ...        [[2, 181, 13, 26, 145, 177, 8, 22, 12, 5, 3755, 1099, 4, 3],
    ...         [2, 46, 145, 500, 1764, 207, 11, 5, 93, 7, 31, 4, 3]]
    ... })
    >>> print(metric.close())
    {'self-bleu': 0.02206768072402293,
     'self-bleu hashvalue': 'c206893c2272af489147b80df306ee703e71d9eb178f6bb06c73cb935f474452',
     'fw-bleu': 0.3831004349785445, 'bw-bleu': 0.025958979254273006, 'fw-bw-bleu': 0.04862323612604027,
     'fw-bw-bleu hashvalue': '530d449a096671d13705e514be13c7ecffafd80deb7519aa7792950a5468549e',
     'gen': [
         ['<go>', 'This', 'is', 'an', 'old', 'photo', 'of', 'people', 'and', 'a', 'horse-drawn', 'wagon', '.'],
         ['<go>', 'An', 'old', 'stone', 'castle', 'tower', 'with', 'a', 'clock', 'on', 'it', '.']
     ]}

Hash value is provided for checking whether the same dataset is used.

note: If you want to know more about metrics, please refer to docs.

Publish Experiments

We provide an online dashboard to manage your experiments.

Here we give an simple example for you.

First initialize a git repo in your command line.

    git init

Then write your model with an entry function in main.py.

    import cotk.dataloader
    import json

    def run():
        dataloader = cotk.dataloader.MSCOCO("resources://MSCOCO_small")
        metric = dataloader.get_inference_metric()
        metric.forward({
            "gen":
                [[2, 181, 13, 26, 145, 177, 8, 22, 12, 5, 3755, 1099, 4, 3],
                [2, 46, 145, 500, 1764, 207, 11, 5, 93, 7, 31, 4, 3]]
        })
        json.dump(metric.close(), open("result.json", 'w'))

note: The only requirement of your model is to output a file named result.json, you can do whatever you want (even don't load data using cotk).

Next, commit your changes and set upstream branch in your command line.

    git add -A
    git commit -a -m "init"
    git remote add origin master https://github.com/USERNAME/REPONAME.git
    git push origin -u master

Finally, type cotk run to run your model and upload to cotk dashboard.

cotk will automatically collect your git repo, username, commit and result.json to the cotk dashboard (TO BE ONLINE).The dashboard is a website where you can manage your experiments or share results with others.

FILL AN IMAGE HERE

If you don't want to use cotk's dashboard, you can also choose to directly upload your model to github.

Use cotk run --only-run instead of cotk run, you will find a .model_config.json is generated. Commit the file and push it to github, the other can automatically download your model as the way described in next section.

note: The reproducibility should be maintained by the author. We only make sure all the input is the same, but difference can be introduced by different random seed, device or other affects. Before you upload, run cotk run --only-run twice and find whether the results is the same.

Reproduce Experiments

You can download others' model in dashboard and try to reproduce their results.

    cotk download ID

The ID comes from dashboard id. cotk will download the codes from dashboard and tell you how to run the models.

INFO: Fetching USERNAME/REPO/COMMIT
13386B [00:00, 54414.25B/s]
INFO: Codes from USERNAME/REPO/COMMIT fetched.
INFO: Model running cmd written in run_model.sh
Model running cmd:  cd ./PATH && cotk run --only-run --entry main

Type cotk run --only-run --entry main will reproduce the same experiments.

You can also download directly from github if the maintainer has set the .model_config.json.

    cotk download USER/REPO/COMMIT

cotk will download the codes from github and generate commands by the config file.

Predefined Models

We have provided some baselines for the classical tasks, see Model Zoo in docs for details.

You can also use cotk download thu-coai/MODEL_NAME/master to get the codes.

Issues

You are welcome to create an issue if you want to request a feature, report a bug or ask a general question.

Contributions

We welcome contributions from community.

• If you want to make a big change, we recommend first creating an issue with your design.
• Small contributions can be directly made by a pull request.
• If you like make contributions for our library, see issues to find what we need.

Team

cotk is maintained and developed by Tsinghua university conversational AI group (THU-coai). Check our main pages (In Chinese).

License

Apache License 2.0