Jupyter Notebooks for Sign Language Translator

Jupyter notebooks using the sign_language_translator package and showing demos as well.

Open in Google Colab

notebooks
├── README.md
├── experiments.ipynb
├── data_collection
│   ├── clip_extractor.ipynb
│   ├── hugging_face_text_datasets.ipynb
│   ├── language_models.ipynb
│   ├── names_LM.ipynb
│   ├── parallel_corpus.ipynb
│   ├── synonyms.ipynb
│   ├── text_cleaning.ipynb
│   ├── text_extraction.ipynb
│   ├── text_scrapper.ipynb
│   ├── trim_clips.ipynb:   analyze position of wrists in frame and remove pauses at start & end
│   └── zip_dataset.ipynb
│
├── model_training
│   ├── embeddings.ipynb
│   ├── ngram_training.ipynb
│   └── transformer_lm_training.ipynb
│
├── text
│   ├── align_embeddings.ipynb
│   ├── frequency.ipynb
│   ├── text_embedding.ipynb
│   └── vocab.ipynb
│
├── translation
│   └── concatenative_synthesis.ipynb:    (text to sign) join sign for each word in text and play
│
└── vision
    ├── concatenate_demo.ipynb
    ├── landmark_display.ipynb
    ├── transforms_demo.ipynb
    ├── video_embedding.ipynb
    ├── video.ipynb
    ├── inputs
    │   └── *
    └── outputs
        └── *