Hugging Face's logo
Activity Feed

AI & ML interests

STARLAB's AI pursues three principles: ethics, transparency, and explainability.

Organization Card
Community About org cards

STARLAB3030

1. AI Model

1.1 Derived from Granite-3.1-8b

Model Description
- ibm-granite/granite-3.1-8b-base IBM Granite Base Model (3.1-8B)
  ├─ starlab3030/rhk_sa_v0.1 RHK SA Model v0.1
  │      └─ starlab3030/rhk_sa_3.1_8b_v1.gguf RHK SA Model v0.1 (8-bit quantized)
  ├─ starlab3030/rhk_sa_v0.2 RHK SA Model v0.2
  └─ starlab3030/rhk_sa_v0.3 RHK SA Model v0.3

2. Skill & Research

2.1 Taxonomy Tree

HowTo

  • Currently, IntructLAB community does not accept the contribution by others
  • Before running SDG, run 'git clone' or copy files to the target taxonomy directory

Sameple Taxonomy Tree for InstructLAB

  • knowledge
    • science/astronomy/constellation/phoenix/qna.yaml is added in Korean version
  • skills
    • Not yet in Korean version

2.2 Markdown

HowTo

  • Only MD & PDF files are allowed
  • Must be published via GitHub

Sample Markdown Files for InstructLAB

  • Markdown
    • phoenix_constellation.md is added in Korean version

datasets 0

None public yet