Skip to content

Kandinsky 6 recipes

All model families

Maintained family · Inference

Kandinsky 6 inference recipes

Kandinsky 6 from the Kandinsky Lab generates five-second video with synchronized audio from text or an image, with base and distilled pi-Flow checkpoints, and upscales existing clips with base or distilled video super-resolution.

Inference live Distillation planned Fine-tuning planned Training planned Evaluation planned Optimization planned Deployment planned

Pick a recipe and runtime

Start with the result you want, then choose one of the runtimes FastVideo actually maintains for it.

Recipe Task and checkpoint
Runtime Maintained paths only

Loading recipe details...

Exact device and memory details appear only when a recorded run supports them.

Reading recipe evidence...

Loading...

Maintained Source-backed Source config
Model
Loading...
Workload
Loading...
Source configuration
Loading...
Expected output
Loading...
Terminal 
Loading...
Open example source View model card

Setup

The generated commands expect a local clone:

git clone https://github.com/hao-ai-lab/FastVideo.git
cd FastVideo

Use Configuration for supported Python and CLI settings, Optimizations for attention and memory tradeoffs, and the support matrix for the supported model and optimization surface.

For model-specific behavior, see Kandinsky 6 video with audio and Kandinsky 6 video super-resolution.

Troubleshooting
  • Text and image conditioning use the same TI2VA pipeline and example. Set IMAGE_PATH in the script to condition on an image; both base and distilled outputs include synchronized audio.
  • For pi-Flow, set KANDINSKY6_MODEL_PATH to the distilled model ID when running basic_kandinsky6_ti2va.py. Its registered preset uses 10 inference steps, guidance 1.0, eps=1e-6, final_step_size_scale=0.5, and num_policy_substeps=128.
  • For VSR, set INPUT_VIDEO in the generated command. The pipeline accepts x2, x2.25 and x4 scales, processes up to 121 frames at 24 fps, and preserves source audio.
  • Checkpoint key-layout errors indicate that the checkpoint and FastVideo checkout target different Kandinsky 6 Diffusers revisions.
  • Gated or missing checkpoints: run huggingface-cli login and confirm you accepted the model's license on Hugging Face.
Evidence status

All recipes on this page are Source-backed: their commands, model IDs, and flags were validated against the checked-in FastVideo sources listed above (static validation). No runtime GPU validation is recorded for these recipes, so GPU model fit, memory use, throughput, and runtime duration are Unknown and deliberately not claimed. Runtime buttons show only the GPU counts configured in checked-in sources.