DakPro

Week 4

---

Repo for the project code. Many of the used services use huggingface client, so setting up huggingface access token is recommended.

---

Setting up access token

---

1. Login in huggingface
2. Goto Settings
3. Goto Access tokens
4. Create a new token (read-only recommended)

Using access token

---

1. brew install huggingface-cli
2. hf auth login
3. Input the access token

When making requests to huggingface client, programs will automatically use the token.

Planned structure of the repo

---

• Outer file transcription_from_mic.py: given a model name runs a runtime transcription demo.
• Outer file transcription_from_file.py: given a model name and file transcribes the file.
  • The irreplaceable part of model pipeline (usually copied from the model source)
• Separate directory for each model, includes
  • Some stuff used before (like reports, scripts)?
  • Interface to use the model, both for demo (with printing captions) and production
• Directory for testing - for interaction with datasets