$100 full-finetune proves MageFlow 4B's potential as an illustration base model
Turbulent-Bass-649 · reddit · 2026-09-07
A GPU-less university student releases MageTrail, a proof-of-concept full-finetune of Microsoft's MageFlow 4B T2I model on a diversity-maximized 41k-image Danbooru/E621 dataset (curated by Chroma creator Lodestone), tuning booru tag prompting and illustration skills in without the $20k-50k+ cost of a full booru tune.
Key points:
- Dataset recaptioned with Gemini 3.7 Flash (backed by Grok 4.6 and Qwen 3.8 27B FP8), updated to 2026 tag standards; everything open-sourced under a Banodoco grant
- V0.1 trained only 30 epochs (300k samples) for $100; learns booru concepts fast but isn't usable yet (lots of seed rerolling for intact limbs)
- Author argues MageFlow 4B's architecture merits investment: 15-20% faster inference than NVIDIA Cosmos2/Anima despite 2B more params, MageVAE beating QwenVAE, native 256-2048px, MIT license, broad knowledge from 10B curated images, minimal forgetting
- Notes Microsoft seemingly purged the model soon after release, and Krea 2 overshadowed it
- Aiming to raise $700 to reach 200 epochs (V0.5) and attract better-funded investment into the arch
More from Multimodal
- FLUX.3 Klein step-distilled to 4-8 steps with CFG baked in — Fit-Structure-6489 · 2026-10-26
- Full animation short made with Google Gemini tools, 10,564 frames — AI_Andrew · 2026-10-26
- Same prompt, two styles: Claude Opus 5.5 vs GPT-6 Astra on a C-section 3D animation — Sourcecode12 · 2026-10-26
- ostris pre-training "Primal VAE": 1-channel latents match modern 8x VAEs via patching — ostrisai · 2026-10-26
- Community checkpoint 'Narrative Drift' brings art-directed visuals to Krea 2 Turbo — MoonbearAIArt · 2026-10-26
- Same prompt, split results: Claude Opus 5.5 vs GPT-6 Astra on a 3D C-section explainer — Sourcecode12 · 2026-10-26