#101 · Primary category: Foundation Models
xlnet
XLNet: Generalized Autoregressive Pretraining for Language Understanding
Project last updated:05/28/23
GitHub Stars
6.2K
Forks
1.1K
Contributors
12
License
Apache-2.0
Why we included this project
XLNet's authors released this repository as the reference implementation of their model, which makes it the place to look for how a strong Transformer-based language model is actually pretrained and fine-tuned rather than just described in a paper. It ships pretrained Base and Large checkpoints with the SentencePiece tokenizer and configs, plus working scripts for text classification, regression, and reading-comprehension tasks such as SQuAD. The permutation-based autoregressive objective departed from the masked denoising approach BERT popularized, and the Transformer-XL backbone helps on tasks where long context matters. Since the code targets TensorFlow 1.x and was built around TPU training, it is not a production serving stack today; treat it as a reference for reproducing experiments and understanding the architecture. Teams wanting to build on or learn from a widely studied pretraining method will find that research value worth the setup.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
transformers
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
CLIP
CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
MiniCPM-V
A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone
generative-models
Generative Models by Stability AI
unilm
Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities