| --- |
| license: apache-2.0 |
| base_model: robbyant/lingbot-video-dense-1.3b |
| pipeline_tag: text-to-video |
| library_name: coreai |
| tags: |
| - coreai |
| - core-ai |
| - coreai-fabric |
| - aimodel |
| - coreml |
| - apple |
| - apple-silicon |
| - on-device |
| - video |
| - text-to-video |
| - image-to-video |
| - vae |
| - diffusion |
| --- |
| |
| > **Canonical:** [`kevinqz/LingBot-Video-Dense-1.3B-CoreAI`](https://huggingface.co/kevinqz/LingBot-Video-Dense-1.3B-CoreAI) — source of truth. |
|
|
| # LingBot-Video Dense 1.3B (fabric) |
|
|
| An Apple Core AI conversion of the **VAE decoder** from |
| [robbyant/lingbot-video-dense-1.3b](https://huggingface.co/robbyant/lingbot-video-dense-1.3b) — the |
| `AutoencoderKLWan` video autoencoder's decode path, mapping a **video latent** to |
| **pixel frames**. Produced by [coreai-fabric](https://github.com/kevinqz/coreai-fabric/blob/main/recipes/lingbot-video-dense-1.3b.yaml) and indexed by |
| [coreai-catalog](https://github.com/kevinqz/coreai-catalog). |
|
|
| > **VAE decoder, not the full video pipeline.** A video diffusion model is four |
| > separable blocks — text encoder, VAE encoder, denoising DiT, and this VAE |
| > decoder. This asset is ONLY the decoder (latent → pixels). The **host owns** the |
| > text encoder, the DiT few-step denoise loop, latent un-normalization |
| > (`latents_mean`/`latents_std`), frame assembly, and — for multi-chunk streaming — |
| > the causal `feat_cache`. It does not, by itself, generate video from a prompt. |
| |
| ## Model facts |
| |
| | Field | Value | |
| |---|---| |
| | Parameters (upstream) | 1.3B | |
| | Architecture | diffusion | |
| | Capabilities | text-to-video | |
| | Deployable core | AutoencoderKLWan VAE decoder (latent → frames) | |
| | Decoded frame shape | 1×3×1×480×832 | |
| | Quantization / precision | none / float32 | |
| | On-disk size | 273 MB | |
| | Asset kind | single-graph VAE decoder (first-chunk decode) | |
| | assetVersion | 2.0 | |
| |
| |
| ## Use it — this needs host code you supply |
| |
| The bundle is a single static-size graph: `z [1,16,T,H/8,W/8]` in → `frames |
| [1,3,Tp,H,W]` out (spatial 8× / temporal 4× upsampling, first-chunk decode). |
| **You supply** the DiT denoise loop that produces the latent, the latent |
| un-normalization, and frame assembly in your host code (Swift or Python). |
| |
| ```bash |
| pip install coreai-catalog && coreai-catalog install lingbot-video-dense-1.3b |
| ``` |
| |
| ## Requirements |
| |
| - **Deployment: macOS 27.0+ / iOS 27.0+, Xcode 27+.** The asset serializes with `minimum_os v27`, |
| so the on-device Swift runtime requires macOS/iOS 27+. A Mac on macOS 26 can |
| convert and inspect it but not run it on-device. |
| - Apple Silicon. |
|
|
| ## Verification (output parity) |
|
|
| - **Gate A (structure): passed** — the bundle's layout + metadata were |
| validated; the graph loads. |
| - **Gate B — graph_output_cosine: 1.000000 min output cosine** (median 1.000000) vs the fp32 torch VAE decoder over 8 seeded latents, measured on apple_silicon. Certifies the export decodes the SAME pixels as the source VAE — a conversion-fidelity metric, not end video quality. |
| - This certifies the export is **numerically faithful to the source VAE decoder** — it |
| does **NOT** certify end video quality. Reproduce with `coreai-fabric verify`. |
| |
| ## Provenance |
| |
| | Field | Value | |
| |---|---| |
| | Base model | [robbyant/lingbot-video-dense-1.3b](https://huggingface.co/robbyant/lingbot-video-dense-1.3b) @ `f9789a7d9b4772a47aba62d4eb5282ddefd1da21` | |
| | Converted by | `models/lingbotvideo/export.py` (version not reported) | |
| | Recipe | [lingbot-video-dense-1.3b](https://github.com/kevinqz/coreai-fabric/blob/main/recipes/lingbot-video-dense-1.3b.yaml) (recipe_source: fabric) | |
| | Precision / quantization | float32 / none | |
| | Conversion date | 2026-07-09 | |
|
|
| Machine-readable, in this repo: |
| [`parity-report.json`](./parity-report.json) · |
| [`reproduce-manifest.json`](./reproduce-manifest.json) · [`LICENSE`](./LICENSE). |
|
|
| ## License and attribution |
|
|
| Weights licensed **apache-2.0** — see the bundled `LICENSE`. This artifact is a **converted derivative** of the base VAE: its |
| weights were converted to Apple Core AI format. The conversion itself is |
| community work. |
|
|
| ## Links |
|
|
| - **Base model:** [robbyant/lingbot-video-dense-1.3b](https://huggingface.co/robbyant/lingbot-video-dense-1.3b) |
| - **Reproduce:** [recipe `lingbot-video-dense-1.3b`](https://github.com/kevinqz/coreai-fabric/blob/main/recipes/lingbot-video-dense-1.3b.yaml) |
| - **Index:** [coreai-catalog](https://github.com/kevinqz/coreai-catalog) |
| - [HF Collection](https://huggingface.co/collections/kevinqz/coreai-apple-on-device-6a4879f21c7e1a87c99bcf5a) |
|
|
|
|
| ## The on-device Core AI ecosystem |
|
|
| - [coreai-fabric](https://github.com/kevinqz/coreai-fabric) — the reproducible |
| recipe → `.aimodel` pipeline that produced this asset. |
| - [coreai-catalog](https://github.com/kevinqz/coreai-catalog) — the index of Core |
| AI models with provenance and integration snippets. |
| - [apple/coreai-models](https://github.com/apple/coreai-models) — Apple's official |
| exporters and runtimes. |
|
|
| ## Not affiliated with Apple |
|
|
| Community conversion. Not produced, hosted, or endorsed by Apple. Apple and Core |
| AI are trademarks of Apple Inc., used here only to describe the target |
| runtime/format. |
|
|
