SDXL LoRA Workflow (Kaggle, reproducible)

This Space summarizes the workflow from mihai-chindris/lora-workflow-template. It is documentation-only and intentionally excludes personal data and weights.

Open workflow repository Open collection

Pipeline steps

  1. Prepare run config and launch training on free Kaggle GPU.
  2. Continue from checkpoints when session limits are reached.
  3. Evaluate checkpoints with template eval scripts.
  4. Generate curated gallery outputs for external presentation.

Included

  • 07_kaggle/train_flux_lora.py
  • 03_configs/* automation/config scripts
  • 08_kaggle_eval/* evaluation templates
  • Runbook notes and reproducibility process

Excluded

  • Personal photos and personal captions
  • Trained LoRA checkpoints or model weights
  • Generated portraits / private galleries
  • Tokens, credentials, and private metadata

Why this Space exists

To provide a clear, auditable workflow artifact you can reference publicly without exposing biometric content, private data, or model outputs tied to personal identity.

Privacy-first publication: share methodology, keep personal data and personal LoRA assets private unless explicitly intended for release.