• • • /projects/streamlined/jsm33/pyannoteIntegration

• Google Drive folder https://drive.google.com/drive/folders/1PshEvfimwzFZTs_IxjY_jVIJKTqkvWmq?usp=sharing

• • Folder: Docker

• • Spreadsheet: Copy of Download ELAR: an exact copy of Emily's spreadsheet, with some updated data links.

• Overleaf
  • Document for ComputEL submission:

Work in Progress

ELAN Integration for speaker diarization models: https://github.com/ngaeghy/ELAN-Diar

Project History (Published Papers)

• • Document for ComputEL submission:

STREANLInED: Shared Tasks for Speech Processing in Language Documentation

Project Background (Published Papers)

Investigating Speaker Diarization of Endangered Language Data

Internal Resources (File Locations)

(Mostly based on Cassie's Handoff Notes, 2022)

• Patas
  • All of the project data and original versions of the LIUM and Kaldi code are found here under /projects/streamlined
  • When you're first granted access, make sure to test out that you can download, edit, access, and upload files, create and move directories, etc... There has been issues in the past because of limited permissions, and it's best to get all sorts of permissions issues out of the way at the beginning of the quarter
  • Main speech data:
    • /projects/streamlined/speech
  • See info in its readme:
    • /projects/streamlined/speech/patas_readme.md (a copy is attached to this page as streamlined_speech.md)
  • Main working directory for Emily (running SID experiments)
    • /projects/streamlined/eahn
    • Emily's bash_profile configs
      • /home2/eahn/.bash_profile
  • Main working directory for Cassie (test running code)
    • /projects/streamlined/cmaz
    • Cassie's bash_profile configs
      • /home2/cassam5/.bash_profile
      • Bash profile configurations will be important for making sure we're all using the same versions of python, kaldi, etc...
  • Directory to run Kaldi SID from:
    • /projects/streamlined/sre_prebuilt/sre08/v1/
  • Minimal java-based pipeline for ELAN integration with pyannote
    • /projects/streamlined/jsm33/pyannoteIntegration
• GitHub
  • Emily Ahn managed the github before. If you need access to private repositories, contact her.
    • Private repo of work in-progress:
      • https://github.com/emilyahn/stream_eahn
      • log.md for weekly notes
    • Public repo of our shared task:
      • https://github.com/emilyahn/streamlined
    • Its corresponding website version:
      • https://emilyahn.github.io/streamlined/
• Docker
  • Login Info
    • Username: cmaz2020
    • Password: #Meetthedockers1
  • The docker profile has working versions of both LIUM and kaldi. Each repository has a description of what is in the container and how to run it.
  • Kaldi container: https://hub.docker.com/repository/docker/cmaz2020/kaldi
  • LIUM/dscore container: https://hub.docker.com/repository/docker/cmaz2020/lium_dscore
  • Summaries of Cassie's work with Docker can be found here:
    • Spring 2021 work (a walkthrough of how to use the Docker containers created up to that point): https://docs.google.com/document/d/1jKXaCteRBrsBb3cmnVdFMplRL7y120p0lKCfEsOj9y8/edit
    • Fall 2021 work (Cassie's progress notes):
      • https://docs.google.com/document/d/1vpSZj6lnr3wZ6U9dUUduimY1Prk0t5tfX-W84KvyYOA/edit
• Google Drive folder https://drive.google.com/drive/folders/1PshEvfimwzFZTs_IxjY_jVIJKTqkvWmq?usp=sharing
  • Folder: Emily_Streamlined
    • This contains Emily Ahn's notes from when she was the project RA. It includes summaries of her work, language info/stats, the original Download ELAR links, and her handoff notes for Cassie (note that some of the links and information may be outdated though)
  • Folder: Docker
    • This contains a walkthrough from Spring 2021, a summary of average kaldi results from the different languages, and a folder of the docker files used for each container (note that the Dockerfile can also be opened when you download and run the container).
  • Spreadsheet: Copy of Download ELAR: an exact copy of Emily's spreadsheet, with some updated data links.
  • Cassie's Handoff Notes including a summary of file locations and where she left off
  • Progress_Notes_Fall_2021: Cassie's stream of consciousness work pipeline for Docker
• Overleaf
  • Document for ComputEL submission (due 10/14):
    • https://www.overleaf.com/project/5f6506df9346310001556491
  • Previous STREAMLInED workshop paper (2016):
    • https://www.overleaf.com/project/5dd77e797bc92700010ca18e

External Resources

ELAN

ELAN Documentations on MPI: https://archive.mpi.nl/tla/elan/documentation

Manual page for Recognizers: https://www.mpi.nl/tools/elan/docs/manual/index.html#Sec_Audio_video_recognizer.html

Other Recognizers:

• XLS-R-ELAN: https://github.com/coxchristopher/xls-r-elan
• Voxseg-ELAN: https://github.com/coxchristopher/voxseg-elan
• Persephone-ELAN: https://github.com/coxchristopher/persephone-elan
• DeepSpeech -ELAN: https://github.com/langdoc/deepspeech-elan

Treehouse Projects

• Grammar Matrix: An open-source starter-kit for the rapid development of cross-lingusitically consistent broad-coverage precision grammars.
• Montage: Software tools which assist field linguists in documenting the grammars of underdescribed languages, while preparing the data and analyses for electronic publication.
• UW TREC '06 Q&A Team: The CLMA program, in collaboration with the iSchool is putting together a system for the TREC 2006 Question Answering System evaluation.

-- EmilyBender - 25 Oct 2004