https://github.com/huggingface/tflite-android-transformers

DistilBERT / GPT-2 for on-device inference thanks to TensorFlow Lite with Android demo apps
https://github.com/huggingface/tflite-android-transformers

android nlp tensorflow tensorflow-lite transformers

Last synced: 2 months ago
JSON representation

DistilBERT / GPT-2 for on-device inference thanks to TensorFlow Lite with Android demo apps

• Host: GitHub
• URL: https://github.com/huggingface/tflite-android-transformers
• Owner: huggingface
• License: apache-2.0
• Created: 2019-11-13T20:47:21.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2023-07-25T06:47:24.000Z (12 months ago)
• Last Synced: 2024-02-02T09:58:57.993Z (5 months ago)
• Topics: android, nlp, tensorflow, tensorflow-lite, transformers
• Language: Java
• Homepage:
• Size: 89 MB
• Stars: 367
• Watchers: 64
• Forks: 75
• Open Issues: 16
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• Awesome-Efficient-Transformers - TensorFlow Lite Transformers w/ Android demos

README

        
# TensorFlow Lite Transformers w/ Android demos

Convert Transformers models

imported from the [🤗 Transformers](https://github.com/huggingface/transformers) library

and use them on Android. You can also check out our

[swift-coreml-transformers](https://github.com/huggingface/swift-coreml-transformers) repo

if you're looking for Transformers on iOS.

# [Question Answering with DistilBERT](bert)

Demo of the [DistilBERT](https://arxiv.org/abs/1910.01108) model

(97% of BERT’s performance on GLUE) fine-tuned for Question answering on the SQuAD dataset.

It provides 48 passages from the dataset for users to choose from.

![demo gif](media/distilbert_qa.gif "Demo running offline on a Samsung Galaxy S8")

# [Text Generation with GPT-2/DistilGPT2](gpt2)

On-device text generation using [GPT-2](https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf) or [DistilGPT2](https://medium.com/huggingface/distilbert-8cf3380435b5) (same distillation process than DistilBERT, 2x faster and 33% smaller than GPT-2)

![demo gif](media/gpt2_generation.gif "Demo running offline on a Samsung Galaxy S8, accelerated")

---

## Models generation

Example scripts used to convert the models are available in the `models_generation` directory.

Please note that they require the nightly version of TensorFlow and might thus be unstable.

## License

[Apache License 2.0](LICENSE)