Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.
Awesome Lists | Featured Topics | Projects
https://github.com/abacaj/awesome-transformers

A curated list of awesome transformer models.
https://github.com/abacaj/awesome-transformers
List: awesome-transformers
Last synced: 8 days ago
JSON representation
A curated list of awesome transformer models.
Host: GitHub
URL: https://github.com/abacaj/awesome-transformers
Owner: abacaj
License: mit
Created: 2023-02-25T04:27:38.000Z (almost 2 years ago)
Default Branch: main
Last Pushed: 2023-04-12T16:20:45.000Z (over 1 year ago)
Last Synced: 2024-12-02T07:48:52.963Z (11 days ago)
Size: 117 KB
Stars: 621
Watchers: 14
Forks: 45
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project

awesome-awesome-artificial-intelligence - Awesome Transformers - transformers?style=social) | (Transformer)
awesome-awesome-artificial-intelligence - Awesome Transformers - transformers?style=social) | (Transformer)
ultimate-awesome - awesome-transformers - A curated list of awesome transformer models. (Other Lists / PowerShell Lists)
README

        # Awesome Transformers

![Transformers](logo.png)

A curated list of awesome transformer models.

If you want to contribute to this list, send a pull request or reach out to me on twitter: [@abacaj](https://twitter.com/abacaj). Let's make this list useful.

There are a number of models available that are not entirely open source (non-commercial, etc), this repository should serve to also make you aware of that. Tracking the original source/company of the model will help.

I would also eventually like to add model use cases. So it is easier for others to find the right one to fine-tune.

_Format_:

- Model name: short description, usually from paper

  - Model link (usually huggingface or github)

  - Paper link

  - Source as company or group

  - Model license

## Table of Contents

- [Encoder (autoencoder) models](#encoder)

  - [ALBERT](#albert)

  - [BERT](#bert)

  - [DistilBERT](#distilbert)

  - [DeBERTaV3](#debertav3)

  - [Electra](#electra)

  - [RoBERTa](#roberta)

- [Decoder (autoregressive) models](#decoder)

  - [BioGPT](#bio-gpt)

  - [CodeGen](#codegen)

  - [LLaMa](#llama)

  - [GPT](#gpt)

  - [GPT-2](#gpt-2)

  - [GPT-J](#gpt-j)

  - [GPT-NEO](#gpt-neo)

  - [GPT-NEOX](#gpt-neox)

  - [NeMo Megatron-GPT](#nemo)

  - [OPT](#opt)

  - [BLOOM](#bloom)

  - [GLM](#glm)

  - [YaLM](#yalm)

- [Encoder+decoder (seq2seq) models](#encoder-decoder)

  - [T5](#t5)

  - [FLAN-T5](#flan-t5)

  - [Code-T5](#code-t5)

  - [Bart](#bart)

  - [Pegasus](#pegasus)

  - [MT5](#mt5)

  - [UL2](#ul2)

  - [FLAN-UL2](#flanul2)

  - [EdgeFormer](#edgeformer)

- [Multimodal models](#multimodal)

  - [Donut](#donut)

  - [LayoutLMv3](#layoutlmv3)

  - [TrOCR](#trocr)

  - [CLIP](#clip)

  - [Unified-IO](#unifiedio)

- [Vision models](#vision)

  - [DiT](#dit)

  - [DETR](#detr)

  - [EfficientFormer](#efficientformer)

- [Audio models](#audio)

  - [Whisper](#whisper)

  - [VALL-E](#valle)

- [Recommendation models](#recommendation)

  - [P5](#p5)

- [Grounded Situation Recognition models](#gsr)

  - [GSRTR](#gsrtr)

  - [CoFormer](#coformer)

  



## Encoder models



- ALBERT: "A Lite" version of BERT

  - [Model](https://huggingface.co/models?other=albert)

  - [Paper](https://arxiv.org/pdf/1909.11942.pdf)

  - Google

  - Apache v2

- BERT: Bidirectional Encoder Representations from Transformers

  

  - [Model](https://huggingface.co/models?other=bert)

  - [Paper](https://arxiv.org/pdf/1810.04805.pdf)

  - Google

  - Apache v2

- DistilBERT: Distilled version of BERT smaller, faster, cheaper and lighter

  

  - [Model](https://huggingface.co/models?other=distilbert)

  - [Paper](https://arxiv.org/pdf/1910.01108.pdf)

  - HuggingFace

  - Apache v2

- DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing

  

  - [Model](https://huggingface.co/models?sort=downloads&search=microsoft%2Fdeberta-v3)

  - [Paper](https://arxiv.org/pdf/2111.09543.pdf)

  - Microsoft

  - MIT

- Electra: Pre-training Text Encoders as Discriminators Rather Than Generators

  

  - [Model](https://huggingface.co/models?other=electra)

  - [Paper](https://arxiv.org/pdf/2003.10555.pdf)

  - Google

  - Apache v2

- RoBERTa: Robustly Optimized BERT Pretraining Approach

  

  - [Model](https://huggingface.co/models?other=roberta)

  - [Paper](https://arxiv.org/pdf/1907.11692.pdf)

  - Facebook

  - MIT



## Decoder models



- BioGPT: Generative Pre-trained Transformer for Biomedical Text Generation and Mining

  - [Model](https://huggingface.co/microsoft/biogpt)

  - [Paper](https://arxiv.org/pdf/2210.10341.pdf)

  - Microsoft

  - MIT

- CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis

  

  - [Model](https://huggingface.co/models?sort=downloads&search=salesforce%2Fcodegen)

  - [Paper](https://arxiv.org/pdf/2203.13474.pdf)

  - Salesforce

  - BSD 3-Clause

- LLaMa: Open and Efficient Foundation Language Models

  

  - [Model](https://github.com/facebookresearch/llama)

  - [Paper](https://research.facebook.com/file/1574548786327032/LLaMA--Open-and-Efficient-Foundation-Language-Models.pdf)

  - Facebook

  - [Requires approval, non-commercial](https://docs.google.com/forms/d/e/1FAIpQLSfqNECQnMkycAp2jP4Z9TFX0cGR4uf7b_fBxjY_OjhJILlKGA/viewform)

- GPT: Improving Language Understanding by Generative Pre-Training

  

  - [Model](https://huggingface.co/openai-gpt)

  - [Paper](https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf)

  - OpenAI

  - MIT

- GPT-2: Language Models are Unsupervised Multitask Learners

  

  - [Model](https://huggingface.co/models?search=gpt-2)

  - [Paper](https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf)

  - OpenAI

  - MIT

- GPT-J: A 6 Billion Parameter Autoregressive Language Model

  

  - [Model](https://huggingface.co/EleutherAI/gpt-j-6B)

  - [Paper](https://github.com/kingoflolz/mesh-transformer-jax)

  - EleutherAI

  - Apache v2

- GPT-NEO: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow

  

  - [Model](https://huggingface.co/models?search=gpt-neo)

  - [Paper](https://doi.org/10.5281/zenodo.5297715)

  - EleutherAI

  - MIT

- GPT-NEOX-20B: An Open-Source Autoregressive Language Model

  

  - [Model](https://huggingface.co/EleutherAI/gpt-neox-20b)

  - [Paper](https://arxiv.org/pdf/2204.06745.pdf)

  - EleutherAI

  - Apache v2

- NeMo Megatron-GPT: Megatron-GPT 20B is a transformer-based language model.

  

  - [Model](https://huggingface.co/nvidia/nemo-megatron-gpt-20B)

  - [Paper](https://arxiv.org/pdf/1909.08053.pdf)

  - NVidia

  - CC BY 4.0

- OPT: Open Pre-trained Transformer Language Models

  

  - [Model](https://huggingface.co/models?search=facebook%2Fopt)

  - [Paper](https://arxiv.org/pdf/2205.01068.pdf?fbclid=IwAR1Fhxr_i3UK3ttigVDGBwbtO-3zLzjTwnyn0dkYt8rf6hxUAUS7Sk7VrYk)

  - Facebook

  - [Requires approval, non-commercial](https://github.com/facebookresearch/metaseq/blob/main/projects/OPT/MODEL_LICENSE.md?fbclid=IwAR2jiCf2R9fTouGGF7v8Tt7Yq8sSVOMot0YIE8ibaP9b2avxw2bEbEaTJZY)

- BLOOM: A 176B-Parameter Open-Access Multilingual Language Model

  

  - [Model](https://huggingface.co/bigscience/bloom)

  - [Paper](https://arxiv.org/pdf/2211.05100.pdf)

  - BigScience

  - [OpenRAIL, use-based restrictions](https://huggingface.co/spaces/bigscience/license)

- GLM: An Open Bilingual Pre-Trained Model

  

  - [Model](https://github.com/THUDM/GLM-130B)

  - [Paper](https://arxiv.org/pdf/2210.02414.pdf)

  - Knowledge Engineering Group (KEG) & Data Mining at Tsinghua University

  - [Custom license, see restrictions](https://github.com/THUDM/GLM-130B/blob/main/MODEL_LICENSE)

- YaLM: Pretrained language model with 100B parameters

  

  - [Model](https://github.com/yandex/YaLM-100B)

  - [Paper](https://medium.com/yandex/yandex-publishes-yalm-100b-its-the-largest-gpt-like-neural-network-in-open-source-d1df53d0e9a6)

  - Yandex

  - Apache v2



## Encoder+decoder (seq2seq) models



- T5: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

  

  - [Model](https://huggingface.co/models?sort=downloads&search=t5)

  - [Paper](https://arxiv.org/pdf/1910.10683.pdf)

  - Google

  - Apache v2

- FLAN-T5: Scaling Instruction-Finetuned Language Models

  

  - [Model](https://huggingface.co/models?sort=downloads&search=flan-t5)

  - [Paper](https://arxiv.org/pdf/2210.11416.pdf)

  - Google

  - Apache v2

- Code-T5: Identifier-aware Unified Pre-trained Encoder-Decoder Models

  for Code Understanding and Generation

  

  - [Model](https://huggingface.co/models?search=code-t5)

  - [Paper](https://arxiv.org/pdf/2109.00859.pdf)

  - Salesforce

  - BSD 3-Clause

- Bart: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension

  

  - [Model](https://huggingface.co/facebook/bart-large)

  - [Paper](https://arxiv.org/pdf/1910.13461.pdf)

  - Facebook

  - Apache v2

- Pegasus: Pre-training with Extracted Gap-sentences for Abstractive Summarization

  

  - [Model](https://huggingface.co/models?sort=downloads&search=pegasus)

  - [Paper](https://arxiv.org/pdf/1912.08777.pdf)

  - Google

  - Apache v2

- MT5: A Massively Multilingual Pre-trained Text-to-Text Transformer

  

  - [Model](https://huggingface.co/models?search=mt5)

  - [Paper](https://arxiv.org/pdf/2010.11934.pdf)

  - Google

  - Apache v2

- UL2: Unifying Language Learning Paradigms

  

  - [Model](https://huggingface.co/google/ul2)

  - [Paper](https://arxiv.org/pdf/2205.05131v1.pdf)

  - Google

  - Apache v2

- FLAN-UL2: A New Open Source Flan 20B with UL2

  

  - [Model](https://github.com/google-research/google-research/tree/master/ul2)

  - [Paper](https://arxiv.org/abs/2205.05131)

  - Google

  - Apache v2

- EdgeFormer: A Parameter-Efficient Transformer for On-Device Seq2seq Generation

  

  - [Model](https://github.com/microsoft/unilm/tree/master/edgelm)

  - [Paper](https://arxiv.org/pdf/2202.07959.pdf)

  - Microsoft

  - MIT



## Multimodal models



- Donut: OCR-free Document Understanding Transformer

  - [Model](https://huggingface.co/models?sort=downloads&search=clova%2Fdonut)

  - [Paper](https://arxiv.org/pdf/2111.15664.pdf)

  - ClovaAI

  - MIT

- LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking

  

  - [Model](https://huggingface.co/models?sort=downloads&search=microsoft%2Flayoutlmv3)

  - [Paper](https://arxiv.org/pdf/2204.08387.pdf)

  - Microsoft

  - CC BY-NC-SA 4.0 (non-commercial)

- TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models

  

  - [Model](https://huggingface.co/models?search=trocr)

  - [Paper](https://arxiv.org/abs/2109.10282)

  - Microsoft

  - [Inherits MIT license](https://github.com/microsoft/unilm/tree/master/trocr#license)

- CLIP: Learning Transferable Visual Models From Natural Language Supervision

  

  - [Model](https://huggingface.co/models?sort=downloads&search=openai%2Fclip)

  - [Paper](https://arxiv.org/pdf/2103.00020.pdf)

  - OpenAI

  - MIT

- Unified-IO: A Unified Model for Vision, Language, and Multi-Modal Tasks

  

  - [Model](https://github.com/allenai/unified-io-inference)

  - [Paper](https://arxiv.org/pdf/2206.08916.pdf)

  - allenai

  - Apache v2



## Vision models



- DiT: Self-supervised Pre-training for Document Image Transformer

  - [Model](https://huggingface.co/models?search=microsoft/dit)

  - [Paper](https://arxiv.org/pdf/2203.02378.pdf)

  - Microsoft

  - [Inherits MIT license](https://github.com/microsoft/unilm/tree/master/dit#license)

- DETR: End-to-End Object Detection with Transformers

  

  - [Model](https://huggingface.co/models?search=facebook/detr)

  - [Paper](https://arxiv.org/pdf/2005.12872.pdf)

  - Facebook

  - Apache v2

- EfficientFormer: Vision Transformers at MobileNet Speed

  

  - [Model](https://huggingface.co/models?sort=downloads&search=snap-research%2Fefficientformer)

  - [Paper](https://arxiv.org/pdf/2206.01191.pdf)

  - Snap

  - Apache v2



## Audio models



- Whisper: Robust Speech Recognition via Large-Scale Weak Supervision

  - [Model](https://huggingface.co/models?sort=downloads&search=openai%2Fwhisper)

  - [Paper](https://arxiv.org/pdf/2212.04356.pdf)

  - OpenAI

  - MIT

- VALL-E: Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers

  

  - [Model (unofficial)](https://github.com/enhuiz/vall-e)

    - MIT but has a dependency on a CC-BY-NC library

  - [Model (unofficial)](https://github.com/lifeiteng/vall-e)

    - Apache v2

  - [Paper](https://arxiv.org/pdf/2301.02111.pdf)

  - Microsoft



## Recommendation models



- Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5)

  - [Model](https://github.com/jeykigung/P5)

  - [Paper](https://arxiv.org/abs/2203.13366)

  - Rutgers

  - Apache v2



## Grounded Situation Recognition models



- Grounded Situation Recognition with Transformers

  - [Model](https://github.com/jhcho99/gsrtr)

  - [Paper](https://arxiv.org/abs/2111.10135)

  - POSTECH

  - Apache v2

- Collaborative Transformers for Grounded Situation Recognition

  

  - [Model](https://github.com/jhcho99/CoFormer)

  - [Paper](https://arxiv.org/abs/2203.16518)

  - POSTECH

  - Apache v2