Peluncuran Stable Diffusion

Stability AI merilis Stable Diffusion — model text-to-image open-source berdasarkan latent diffusion model. Demokratisasi AI image generation.

Stable Diffusion 1.4: open-source (CreativeML Open RAIL-M), 890M parameters, berjalan di GPU consumer (8GB VRAM). Pendorong boom Civitai, Hugging Face diffusion.

Print

Peluncuran Stable Diffusion

Ringkasan

Stability AI merilis Stable Diffusion 1.4 pada 22 Agustus 2022 — model text-to-image open-source berdasarkan latent diffusion model dari CompVis Group (LMU Munich) dan RunwayML.

Signifikansi

  • Open-source — lisensi CreativeML Open RAIL-M (gratis untuk komersial)
  • Consumer GPU — berjalan di GPU 8GB VRAM (RTX 3060+)
  • Demokratisasi — model sebanding DALL-E 2 ada di tangan siapa saja

Komunitas

  • Civitai — marketplace model & LoRA
  • Hugging Face — model repository
  • AUTOMATIC1111 — WebUI (paling populer)
  • ComfyUI — node-based workflow
  • InvokeAI — alternatif UI
  • ControlNet (2023) — struktur kontrol
  • LoRA (2023) — efficient fine-tuning
  • Stable Diffusion XL (2023) — 2.6B parameters
  • SD3, SD3.5 (2024) — MMDiT

Dampak

  • Open-source generative AI sebagai gerakan
  • $2.6 miliar pendanaan Stability AI (peak 2023)
  • Pertumbuhan Civitai dari 0 ke 5M+ user (2023)
  • Komunitas — Dreambooth, LoRA, Textual Inversion
  • Kontroversi — deepfake NSFW, seni style cloning

Stable Diffusion adalah counter-budaya AI — alternatif open-source terhadap proprietary model.

Connected to

Not yet written

The following pages are referenced but don't exist yet — they'd make good future additions.

  • /concepts/diffusion-model
  • /sources/stability-ai

References

  1. Wikipedia

Type at least 2 characters to search.

Press to navigate, to open, esc to close.