README.md

metadata

license: apache-2.0
tags:
  - scripts
  - research
  - training

Azure scripts

Miscellaneous scripts from the Azure H100 workspace. Not a deployable model — this is a working collection.

Contents

• train_h100_*.py, train_v3.py — training runs (various projects)
• generate_dataset*.py — data generation
• close_the_loop.py — TRIBE v2 AI↔brain comparison scaffolding
• merge_v3.py, pull_extra.py — dataset maintenance
• telugu_voice_clone.py — voice synthesis experiment
• yc_scrape.py — scraping
• gsoc_proposal_*.md, GSoC_2026_Proposal_Vivek_OLMo.docx — DeepChem GSoC 2026 proposal (OLMo wrapper)
• turboquant_case_study.md — TurboQuant KV cache paper implementation notes
• vivek_complete_profile.md
• h100_training.log, train_*.log — historical training logs
• AI4RTGSdevDB.camera_masters.json — camera master config (schema only)