veizik WAN (CUDA) v1.2.93 — Wan 2.1 & Wan 2.2 video, warm hypercache — cuda-wan-1.2.93

Text-to-video and image-to-video with Wan2.1-T2V-1.3B, Wan2.1-T2V-14B, Wan2.2-T2V-A14B and Wan2.2-I2V-A14B on your own NVIDIA GPU, through veizik's own native engine: no Python, no CUDA toolkit. You supply the weights; the frames are written as PNG files to the folder you name. This release adds an optional warm model-load — a sealed per-model hypercache you can pull from the CDN so a model starts from a prepared snapshot instead of building cold, with an automatic cold-build fallback when it is absent. It also runs a resident render server with a host-loop (bin/veizik-resident) that keeps one process and one GPU context across requests, and mid-render preemption. Distributed as a single signed .vzpack, and render-verified fresh on the shipped bytes (RTX 3090, sm_86).

This page records the state as of 2026-11-14 and is not updated as the catalogue moves. For what runs now, see models.

What this build does

• veizik run <model key> "prompt" --out DIR renders a video and writes its frames as PNG files to DIR/frames/; nothing is written inside the package.
• Serve: veizik serve <model> runs a resident render server that renders request after request in one process and one GPU context, with results bit-identical to veizik run.
• Four modes: quality (the reference path and the default), balanced, fast and ultrafast.
• An interrupted render continues from its last step with --resume, and the result is identical to a render that never stopped.
• Every file of the package is checked against its sealed manifest before each run.
• GPU code for sm_80, sm_86, sm_89, sm_90a, sm_100a and sm_120a.
• Every frame it writes carries the provenance mark described in the Terms: an in-pixel watermark plus XMP DigitalSourceType trainedAlgorithmicMedia.
• Distributed as a single signed .vzpack: veizik install wan22 fetches and verifies it (and the shared-core .vzroot once) from its GitHub release and refuses any other file.
• mid-render preemption — POST /v1/videos/generations with "priority": "urgent" interrupts a running render at a step boundary, is served, then the first render resumes from its checkpoint.
• Optional warm model-load: veizik pull --hypercache <model> fetches a sealed per-model hypercache from the CDN so the model starts from a prepared snapshot instead of building cold; if it is absent the engine falls back to a cold build. The hypercache is covered by the package licence — no separate key.

Models in this release

Platform

Linux on x86-64

Verification

`veizik install` checks the download against SHA-256 c35b003d8a0a493b5b79904dd28bd1b793293061383fe6de3c498435dc5c6087 and refuses any other file. The release is reproducible: SHA256SUMS on the release, a single-file .vzpack with a signed .vzindex. To check veizik-wan-1.2.93-linux-x86_64.vzpack by hand, compare it with sha256sum.

File

veizik-wan-1.2.93-linux-x86_64.vzpack

Size

7,166,271 bytes

SHA-256

c35b003d8a0a493b5b79904dd28bd1b793293061383fe6de3c498435dc5c6087

GPU architectures

sm_80, sm_86, sm_89, sm_90a, sm_100a, sm_120a

Run end to end on

GeForce RTX 3090 (sm_86)

Release

v1.2.93

Get, install, activate

# the veizik command for Linux x86-64
$ curl -fsSL https://veizik.com/install.sh | sh
# if the installer asks you to put ~/.local/bin on your PATH
$ export PATH="$HOME/.local/bin:$PATH"
# accepted names: wan21, wan22 or a full model key (wan21_t2v_1_3b, wan21_t2v_14b, wan22_t2v_a14b, wan22_i2v_a14b, wan22_i2v_a14b_lightning); any case; '.' and '-' work too (wan2.2)
$ veizik install wan22
$ veizik activate

Run `veizik activate` once per device after `veizik install`. Then `veizik pull wan22` fetches the Wan weights and prepares the video decoder for you — no manual VAE setup or support request. Activation itself does not need the weights; rendering does. veizik install refuses a download whose SHA-256 differs from this release. The other packages are on the download page.

Fixed issues

• Adds optional warm model-load (a sealed per-model hypercache) for the four Wan models, verified by a fresh-customer end-to-end run on the shipped bytes: install, signed catalogue present, pull --hypercache, run, and the output matches the reference oracle.
• A pre-release package that did not embed the signed catalogue — a fresh customer's pull --hypercache returned 'no signed catalogue' — was caught before launch and fixed by the catalogue-embedded re-cut.
• Render-verified fresh on the shipped bytes by our own testing (RTX 3090, sm_86).

Known issues in cuda-wan-1.2.93

• Model weights are not included. veizik pull wan22 downloads the Wan2.1/2.2 weights (Apache-2.0) and prepares the video decoder automatically — no manual VAE setup; veizik pull --prep-local DIR prepares weights you already have.
• Speed figures for this package were measured on an NVIDIA RTX 3090 (sm_86) only.
• The image-to-video model takes its input image as raw RGB bytes, one byte per channel, at exactly the render size (width × height × 3 bytes).
• balanced, fast and ultrafast are not bit-identical to quality; without --det-vae yes decoded frames may differ slightly from run to run.
• A distilled image-to-video Lightning model is render-verified for this build and is being enabled on the site with its own benchmark; it is not yet listed as a separate model key.