This model requires you to accept the Apache 2.0 license terms before download.
Description
Text-to-video generation at native 720x1280, 88-frame clips, 3D-RoPE diffusion transformer with T5 conditioning.
Tags
At a glance
| Kind | video generation model |
|---|---|
| Precision | bfloat16 |
| Container size | 23.6 GiB |
| Weights inside | 23.6 GiB |
| Components | 3 |
| Licence | Apache 2.0 |
What is inside the container
A .nbx container packs every part of Allegro — 3 components with weights, graphs and runtime settings — so one file is all the runtime needs.
| Component | Blocks | Layers | Weights |
|---|---|---|---|
| transformer | 32 | 32 | 5.2 GiB |
| vae | 4 | — | 667 MiB |
| text_encoder | 24 | 24 | 17.7 GiB |
Default settings
The values the container ships with. They apply when you do not pass your own.
- Frames
- 88
- Frames per second
- 15
- Inference steps
- 100
- Guidance scale
- 7.5
Memory needed
The weights of Allegro take 23.6 GiB in bfloat16. That is the floor for the memory the runtime has to place, spread over the devices it finds; activations come on top and grow with the size and length you ask for. A measured peak per device class is not published for this model yet.
Verify what you downloaded
Every download carries its SHA-256, and the file endpoint sends it in the X-NBX-SHA256 header. The container is 23.6 GiB (25,294,598,292 bytes).
$ sha256sum Allegro.nbx
f59950361bfa9563be2c5f1e6797b6b450fb66e7205d372bcd067de9a9488c8b
Questions
What is Allegro?
A video generation model by rhymes-ai, distributed as a bfloat16 .nbx container and licensed Apache 2.0.
How big is the download?
23.6 GiB, of which 23.6 GiB are weights.
How do I run it?
Install the runtime from the download page, then use the commands under “CLI Usage” below: neurobrix import rhymes-ai/Allegro, then neurobrix serve --model Allegro.
CLI Usage
$ neurobrix import rhymes-ai/Allegro
$ neurobrix serve --model Allegro
$ neurobrix run --prompt "your prompt here"
$ neurobrix stop