Skip to main content
skill CX Codex

anymove-lora — Claude Code skill: what it does, how to install

anymove-lora is a Claude Code skill that trains a MiniMax H3 LoRA on fal to preserve a creature's two-camera frame and limb colors.

Apache-2.0 ★ 653 updated 2026-11-13

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add patch-gg/anymove
/plugin install anymove@anymove
Open on GitHub

What anymove-lora does

anymove-lora trains and runs a MiniMax H3 image-to-video LoRA on fal. It keeps the split-screen layout, matching views, and limb colors when plain video generation breaks that arrangement.

The dataset starts with clean captured motions. Blender renders both layout cameras, mirrored clips are added, captions are attached, and the set is packed before a debug run, full training, and an A/B comparison with plain H3 using matching seeds. The debug pass costs $0.01; 1,200 training steps cost about $12 and take about 2 h.

The skill needs FAL_KEY, Blender 5.x, and a creature with several good captures from anymove-capture. Training runs on fal rather than locally. Mirroring an asymmetric rig, such as dragon wings, needs checking. MiniMax H3 and fal have their own licence terms.

What it helps with

  • preserve a two-camera split screen in generated video
  • keep limb colors from swapping between views
  • maintain sync across both views of one motion
  • train a LoRA from captured and mirrored clips
  • compare the LoRA with plain H3 on matching seeds

Example prompts for the agent

  • Train a LoRA for the creature using captures in runs/<c>/clips/, then inspect both views in the rendered frames.
  • Build the wolf dataset, run the debug pass, and inspect runs/wolf/lora/debug/debug_dataset.zip.
  • Generate a held-out pounce video with the LoRA and compare it with plain H3 using the same seed.

How to install anymove-lora

1Plugin marketplace

/plugin marketplace add patch-gg/anymove
/plugin install anymove@anymove

This installs the whole anymove plugin with all its skills

2Skills CLI (Vercel)

npx skills add patch-gg/anymove --skill anymove-lora -a claude-code

3Manually: put the files in place

~/.claude/skills/anymove-lora/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/anymove-lora/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about anymove-lora

How do I install anymove-lora in Claude Code?

Add the marketplace with /plugin marketplace add patch-gg/anymove, then install the plugin with /plugin install anymove@anymove.

What does anymove-lora do?

It trains a MiniMax H3 image-to-video LoRA on fal from two-camera motion clips and their mirrored versions. It then compares the result with plain H3 using matching seeds.

What does anymove-lora need to run?

It needs a FAL_KEY, Blender 5.x, and a creature with several good captures from anymove-capture. Training runs on fal rather than on the local machine.

Source text

SKILL.md · Apache-2.0
---
name: anymove-lora
description: Train and use a MiniMax H3 image-to-video LoRA on fal that keeps anymove's two-camera layout - split screen held, limb colours kept, both views in step. Builds the dataset by re-rendering clean captured motions (and their mirrors) from both cameras, packs it, runs a cheap debug pass, trains, and A/B tests against plain H3 with the same seeds. Use when plain video models keep swapping limb colours or breaking the split screen for a creature, or when someone asks about @Graalitoo's LoRA step, its cost or its training data.
license: Apache-2.0
compatibility: Needs a fal.ai key (FAL_KEY), Blender 5.x, and a creature with several good captures (anymove-capture). Training runs on fal, not locally.
---

# anymove-lora: teach the video model the layout

Graalitoo's addition to the method: a LoRA on MiniMax H3 (open weights) so two-camera clips come out right more
often. fal hosts both the trainer (`minimax/h3/i2v/trainer`, $0.01 a step) and LoRA inference
(`minimax/h3/image-to-video/lora`, about $0.375 a 5 s clip, about 200 s each). No local GPU needed for this part.

## The data, and what it is

Each training clip is one motion the pipeline already captured cleanly (paw agreement 0.8 or more), re-rendered in
Blender from both layout cameras on the painted rig, composed into the two-camera frame, and captioned with the prompt
that produced it. Each is also mirrored left to right (left/right and colour words swapped in the action sentence), so
19 motions give 38 clips. Every clip comes from one 3D motion, so the two halves are in sync by construction. The
motion itself is AI-made: it was solved from AI clips. Real animal mocap exists (for dogs, AI4Animation's data) but
its licence is non-commercial; say which you're using.

## Steps

1. Capture 15-20 good moves (anymove-film, anymove-capture). Prompts live in `prompts/<c>/<move>.txt` or
   `prompts/<c>/train/<move>.txt`; the wolf has 14 training and 4 held-out moves already written.
2. Build the set (Blender renders each clip and its mirror, 5 at a time):
   ```bash
   MOCAP_CREATURE=assets/<c> .venv/bin/python tools/build_lora_set.py runs/<c>/lora/data trot_s1 howl_s1 ...
   ```
   Clip names are folders under `runs/<c>/clips/`; the move is the name up to its last `_`. It was built and tested on
   the wolf, whose rig is symmetric; mirroring an asymmetric rig (the dragon's wings) needs checking.
3. Pack, then a 1-step debug run ($0.01) to see what fal's preprocessing does (768 x 448, 107 frames at 24 fps):
   ```bash
   .venv/bin/python -m mocap.lora pack runs/<c>/lora/data runs/<c>/lora/data.zip
   .venv/bin/python -m mocap.lora train runs/<c>/lora/data.zip runs/<c>/lora/debug 1 --debug
   ```
   Open `runs/<c>/lora/debug/debug_dataset.zip` and check the frames still show both views whole.
4. Train. Say the cost first: 1,200 steps is about $12 and about 2 h (rank 32, lr 2e-4, trigger word `TWOCAM`).
   ```bash
   setsid .venv/bin/python -m mocap.lora train runs/<c>/lora/data.zip runs/<c>/lora/v1 1200 > runs/<c>/lora/v1.log 2>&1 &
   ```
   If the session drops, the run keeps going on fal: `python -m mocap.lora resume runs/<c>/lora/v1` re-attaches.
5. Generate with it (the trigger word is added for you):
   ```bash
   .venv/bin/python -m mocap.lora generate runs/<c>/lora/v1 runs/<c>/ref/ref.png prompts/<c>/pounce.txt runs/<c>/videos/pounce/lora.mp4 1.0 42
   ```
   The wolf's trained LoRA is in the examples release, so it can be tried without training (about $0.38 a clip);
   fal loads it straight from the release:
   ```bash
   tools/fetch_examples.sh wolf --clips
   .venv/bin/python -m mocap.lora generate \
     https://github.com/Patch-gg/anymove/releases/download/examples-v1/lora-wolf-twocam-h3.safetensors \
     runs/wolf/ref/ref.png prompts/wolf/heldout/pounce.txt runs/wolf/videos/pounce/lora.mp4 1.0 42
   ```
   `tools/fetch_examples.sh wolf --lora` downloads the file itself (131 MB), to run H3 with it yourself.
6. A/B test on held-out moves against plain H3 with the same seed, then capture both and compare paw agreement and
   reprojection error (anymove-capture).

## What it did for the wolf (one seed per move: a first read, not a benchmark)

| move | plain H3 | H3 + LoRA |
|---|---|---|
| trot (in training set) | colours swapped, paws 0.15, fit 20.5 px | colours kept, paws 0.48, fit 6.0 px |
| gallop | colours swapped, fit 44 px | colours kept, paws -0.06, fit 21.6 px |
| pounce | one paw moves (-0.65), fit 6.0 px | all four move, paws 0.80, fit 6.2 px |
| dig | paws 0.75, fit 5.6 px | paws 0.87, fit 4.8 px |
| shake | paws 0.70, fit 9.1 px | paws 0.69, fit 6.9 px |

Plain H3 broke the colour code on 2 of 5 clips, the LoRA on 0 of 5. Sync improved on trot, pounce and dig, not on the
gallop. MiniMax H3's open weights carry a community licence (its page asks US, EU, UK and South Korea users to apply);
hosted use on fal goes through fal's terms. Check both before shipping anything made with it.

© Patch · Apache-2.0 · patch-gg/anymove

Similar skills

All skills for AI and ML →

From the same set7

Project sponsors

Companies that keep this analytics open