Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/LogiTorch/logitorch

LogiTorch is a PyTorch-based library for logical reasoning on natural language
https://github.com/LogiTorch/logitorch

Last synced: 5 days ago
JSON representation

LogiTorch is a PyTorch-based library for logical reasoning on natural language

Host: GitHub
URL: https://github.com/LogiTorch/logitorch
Owner: LogiTorch
License: apache-2.0
Created: 2021-08-11T14:56:38.000Z (about 3 years ago)
Default Branch: main
Last Pushed: 2024-09-10T18:52:00.000Z (about 2 months ago)
Last Synced: 2024-09-20T06:51:53.068Z (about 2 months ago)
Language: Python
Homepage: https://logitorch.ai
Size: 530 KB
Stars: 66
Watchers: 4
Forks: 5
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

Awesome-LLM-Reasoning - LogiTorch - based library for logical reasoning on natural language. (Other Useful Resources / 2023)

README

        # LogiTorch

![image](docs/source/_static/logo_with_text.jpg)

[![Downloads](https://static.pepy.tech/badge/logitorch)](https://pepy.tech/project/logitorch)

LogiTorch is a PyTorch-based library for logical reasoning on natural language, it consists of:

- Textual logical reasoning datasets

- Implementations of different logical reasoning neural architectures

- A simple and clean API that can be used with PyTorch Lightning

## 📦 Installation

```console

foo@bar:~$ pip install logitorch==0.0.1a2

```

Or

```console

foo@bar:~$  pip install git+https://github.com/LogiTorch/logitorch.git

```

## 📖 Documentation

You can find the documentation for LogiTorch on [ReadTheDocs](https://logitorch.readthedocs.io).

## 🖥️ Features

### 📋 Datasets

Datasets implemented in LogiTorch:

- [x] [AR-LSAT](https://arxiv.org/abs/2104.06598) _{(MIT LICENSE)}

- [x] [ConTRoL](https://arxiv.org/abs/2011.04864) _{(GitHub LICENSE)}

- [x] [LogiQA](https://arxiv.org/abs/2007.08124) _{(GitHub LICENSE)}

- [x] [ReClor](https://arxiv.org/abs/2002.04326) _{(Non-Commercial Research Use)}

- [x] [RuleTaker](https://arxiv.org/abs/2002.05867) _{(APACHE-2.0 LICENSE)}

- [x] [ProofWriter](https://arxiv.org/abs/2012.13048) _{(APACHE-2.0 LICENSE)}

- [x] [SNLI](https://arxiv.org/abs/1508.05326) _{(CC-BY-SA-4.0 LICENSE)}

- [x] [MultiNLI](https://arxiv.org/abs/1704.05426) _{(CC-BY-SA-4.0 LICENSE)}

- [x] [RTE](https://tac.nist.gov/publications/2010/additional.papers/RTE6_overview.proceedings.pdf) _{([TAC User Agreements](https://tac.nist.gov//data/forms/index.html))}

- [x] [Negated SNLI](https://aclanthology.org/2020.emnlp-main.732/) _{(MIT LICENSE)}

- [x] [Negated MultiNLI](https://aclanthology.org/2020.emnlp-main.732/) _{(MIT LICENSE)}

- [x] [Negated RTE](https://aclanthology.org/2020.emnlp-main.732/) _{(MIT LICENSE)}

- [x] [PARARULES Plus](https://github.com/Strong-AI-Lab/PARARULE-Plus) _{(MIT LICENSE)}

- [x] [AbductionRules](https://arxiv.org/abs/2203.12186) _{(MIT LICENSE)}

- [x] [FOLIO](https://arxiv.org/abs/2209.00840) _{(CC-BY-SA-4.0 LICENSE)}

- [x] [FLD](https://proceedings.mlr.press/v202/morishita23a.html)  _{(CC-BY-SA-4.0 LICENSE)}

- [x] [LogiQA2.0](https://arxiv.org/abs/2007.08124) _{(CC-BY-SA-4.0 LICENSE)}

- [ ] [LogiQA2.0 NLI](https://arxiv.org/abs/2007.08124) 

- [ ] [HELP](https://aclanthology.org/S19-1027.pdf) 

- [ ] [SimpleLogic](https://arxiv.org/abs/2205.11502)

- [ ] [RobustLR](https://arxiv.org/abs/2205.12598)

- [ ] [LogicNLI](https://aclanthology.org/2021.emnlp-main.303/)

### 🤖 Models

Models implemented in LogiTorch:

- [x]  [RuleTaker](https://arxiv.org/abs/2002.05867)

- [x]  [ProofWriter](https://arxiv.org/abs/2012.13048)

- [x]  [BERTNOT](https://arxiv.org/abs/2105.03519)

- [x]  [PRover](https://arxiv.org/abs/2010.02830)

- [x]  [FLDProver](https://proceedings.mlr.press/v202/morishita23a.html)

- [ ]  [TINA](https://suchanek.name/work/publications/emnlp-2022.pdf)

- [ ]  [FaiRR](https://arxiv.org/abs/2203.10261)

- [ ]  [LReasoner](https://arxiv.org/abs/2105.03659)

- [ ]  [DAGN](https://arxiv.org/abs/2103.14349)

- [ ]  [Focal Reasoner](https://arxiv.org/abs/2105.10334)

- [ ]  [AdaLoGN](https://arxiv.org/abs/2203.08992)

- [ ]  [Logiformer](https://arxiv.org/abs/2205.00731)

- [ ]  [LogiGAN](https://arxiv.org/abs/2205.08794)

- [ ]  [MERit](https://arxiv.org/abs/2203.00357)

- [ ]  [APOLLO](https://arxiv.org/abs/2212.09282)

- [ ]  [LAMBADA](https://arxiv.org/abs/2212.13894)

- [ ]  [Chainformer](https://aclanthology.org/2023.findings-acl.588/)

- [ ]  [IDOL](https://aclanthology.org/2023.findings-acl.513/)

## 🧪 Example Usage

### Training Example

```python

import pytorch_lightning as pl

from pytorch_lightning.callbacks import ModelCheckpoint

from torch.utils.data.dataloader import DataLoader

from logitorch.data_collators.ruletaker_collator import RuleTakerCollator

from logitorch.datasets.qa.ruletaker_dataset import RuleTakerDataset

from logitorch.pl_models.ruletaker import PLRuleTaker

train_dataset = RuleTakerDataset("depth-5", "train")

val_dataset = RuleTakerDataset("depth-5", "val")

ruletaker_collate_fn = RuleTakerCollator()

train_dataloader = DataLoader(

    train_dataset, batch_size=32, collate_fn=ruletaker_collate_fn

)

val_dataloader = DataLoader(

    val_dataset, batch_size=32, collate_fn=ruletaker_collate_fn

)

model = PLRuleTaker(learning_rate=1e-5, weight_decay=0.1)

checkpoint_callback = ModelCheckpoint(

    save_top_k=1,

    monitor="val_loss",

    mode="min",

    dirpath="models/",

    filename="best_ruletaker",

)

trainer = pl.Trainer(callbacks=[checkpoint_callback], accelerator="gpu", gpus=1)

trainer.fit(model, train_dataloader, val_dataloader)

```

### Pipeline Example

We provided pre-configured pipelines for some datasets to train models.

```python

from logitorch.pipelines.qa_pipelines import ruletaker_pipeline

from logitorch.pl_models.ruletaker import PLRuleTaker

model = PLRuleTaker(learning_rate=1e-5, weight_decay=0.1)

ruletaker_pipeline(

    model=model,

    dataset_name="depth-5",

    saved_model_name="models/",

    saved_model_path="best_ruletaker",

    batch_size=32,

    epochs=10,

    accelerator="gpu",

    gpus=1,

)

```

### Testing Example

```python

from logitorch.pl_models.ruletaker import PLRuleTaker

from logitorch.datasets.qa.ruletaker_dataset import RULETAKER_ID_TO_LABEL

model = PLRuleTaker.load_from_checkpoint("models/best_ruletaker.ckpt")

context = "Bob is smart. If someone is smart then he is kind."

question = "Bob is kind."

pred = model.predict(context, question)

print(RULETAKER_ID_TO_LABEL[pred])

```

## Citing

Users of LogiTorch should distinguish the datasets and models of our library from the originals. They should always credit and cite both our library and the original data source, as in ``We used LogiTorch's \cite{helwe2022logitorch} re-implementation of BERTNOT \cite{hosseini2021understanding}''.

If you want to cite LogiTorch, please refer to the publication in the [Empirical Methods in Natural Language Processing](https://2022.emnlp.org/):

```code

@inproceedings{helwe2022logitorch,

  title={LogiTorch: A PyTorch-based library for logical reasoning on natural language},

  author={Helwe, Chadi and Clavel, Chlo\'e and Suchanek, Fabian},

  booktitle={Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations},

  year={2022}

}

```

## Acknowledgments

This work was partially funded by ANR-20-CHIA-0012-01 (“NoRDF”).