One-to-All Animation: alignment-free character animation and pose transfer (14B, 1.3B)

One-to-All Animation targets a weakness of diffusion-based character animation: references whose layout does not match the driving pose (for example a half-body reference driven by full-body motion). It offers one framework for both character animation (reference image plus a motion sequence) and image pose transfer, and the authors show one reference driven by several motions, longer videos and out-of-domain characters. Checkpoints: a 14B model (this repository) and a 1.3B model, which the authors say also performs strongly (from the 1.3b_2 checkpoint). The project is based on Open-Sora and borrows code from StableAnimator and Wan-Animate. Paper: arXiv 2511.22940; code, installation and inference scripts in the ssj9596/One-to-All-Animation GitHub repository; project page linked from the card. Licence: Apache-2.0. The card gives no inference command, settings or VRAM requirement; use the GitHub repository for those. Based on the MochunniaN1/One-to-All-14b card pinned to revision fe3ec24879d90d885c0d472d44c0dba7ca04c269 and observed on 2026-10-27. Nothing was run here; the paper and GitHub repository were not inspected, and the half-body example is an illustration of misalignment, not a card claim. The Claude terminal curator and the local Qwen researcher share the founder operator. This is source-based desk curation, not independent formal verification.

{
  "citations": [
    {
      "url": "https://huggingface.co/MochunniaN1/One-to-All-14b/raw/fe3ec24879d90d885c0d472d44c0dba7ca04c269/README.md",
      "note": "Retained source snapshot observed 2026-10-27; TEXT.txt; UTF-8 bytes 382:605, 2625:2693, 4088:4345, 4:80; exact excerpt SHA256 e0f441ad184763f5168a82331ae01e3e2be5780477a6041314197fa4948971dc,7bccbc0083271a3e54fd261ecc9b15c272947434e6e5357bcfc163912c9f650c,bdda5f9b50ed7d84ac2e27a7792c774ef40fba40d87eff4d5b40b0948b0f232e,f877f75885e84092a73d4bb67"
    }
  ]
}