Hugging Face's logo

kandinskylab
/
Kandinsky-6.0-Lite-5s-Diffusers

Image-to-Video
Diffusers
Safetensors
Kandinsky6TI2VAPipeline
video
video-generation
text-to-video
audio-generation
kandinsky
Model card Files
xet
 Community
1
Kandinsky-6.0-Lite-5s-Diffusers

Commit History

27 days ago

  1. Update README.md

    leffff
    Verified signature
    8ec0176

  1. Update README.md

    KirKoz
    Verified signature
    5686616

  1. Link Diffusers and vLLM badges

    leffff
    Verified signature
    50029a5

  2. Update README.md

    leffff
    Verified signature
    c276fdb

  3. Model card: HF Demo badge links to the Space, Report badge links to the tech report⁠

    leffff +1: Claude Sonnet 5
    6510114

  4. Model card: larger logo (80% width)⁠

    leffff +1: Claude Sonnet 5
    777f949

  5. Model card: show the logo with a plain img instead of picture, so it renders⁠

    leffff +1: Claude Sonnet 5
    48be5c7

  6. Model card: comment out the SGLang, vLLM and FastVideo badges⁠

    leffff +1: Claude Sonnet 5
    9072317

  7. Upload promo.webp

    leffff
    Verified signature
    3e9b5ee

  8. Model card: Super Res example uses the flow-matching VSR-5s checkpoint (4 steps) with flex backend and compile_repeated_blocks⁠

    leffff +1: Claude Sonnet 5
    a25c7ee

  1. Model card: Super Res example uses the distilled VSR checkpoint (2 steps) with inductor autotune⁠

    leffff +1: Claude Sonnet 5
    beac12b

  2. Model card: trim usage and super-resolution sections, drop RL-tuned wording⁠

    leffff +1: Claude Sonnet 5
    7de1d1b

  3. Model card: absolute image URLs so the logo renders; TI2AV prompt matches the reference image⁠

    leffff +1: Claude Sonnet 5
    4b2ae52

  4. Model card: default T2AV prompt and image-to-audio-video example⁠

    leffff +1: Claude Sonnet 5
    fe63e08

  5. Update model card: common Kandinsky 6.0 header, badges, usage and super-resolution guides⁠

    leffff +1: Claude Sonnet 5
    d03bcdd

  1. Re-key transformer state_dict for the FeedForward/TimestepEmbedding/attn/dec_block/private_/ModuleList/OutputHead module renames

    leffff
    Verified signature
    5e7cbb7

  1. Add usage examples, upscale pipeline and performance/optimization docs to the model card

    leffff
    Verified signature
    c340bb2

  1. Add split vocoder component (audio_vae unchanged, still usable by sglang's combined loader)

    leffff
    Verified signature
    09c1317

  1. Add transformer (DiT) weights: 3B TI2V SFT checkpoint

    leffff
    Verified signature
    f0dd5c4

  2. Copy shared components from Pro-5s (excluding transformer)

    leffff
    Verified signature
    f94362e

  1. initial commit

    leffff
    Verified signature
    3a6098c