This model requires you to accept the Apache 2.0 license terms before download.
Description
Flex.1 Alpha — Ostris 8-billion-parameter rectified-flow text-to-image model derived from FLUX.1-schnell, featuring a trainable guidance embedder that makes it highly flexible to fine-tune. Apache 2.0.
Tags
At a glance
| Kind | image generation model |
|---|---|
| Precision | bfloat16 |
| Container size | 24.5 GiB |
| Weights inside | 24.5 GiB |
| Components | 4 |
| Licence | Apache 2.0 |
| Origin | ostris/Flex.1-alpha @ 5dd0b1bc2a94 |
What is inside the container
A .nbx container packs every part of Flex.1-alpha — 4 components with weights, graphs and runtime settings — so one file is all the runtime needs.
| Component | Blocks | Layers | Weights |
|---|---|---|---|
| transformer | 38 | 8 | 15.2 GiB |
| vae | 4 | — | 160 MiB |
| text_encoder | 12 | — | 235 MiB |
| text_encoder_2 | 24 | 24 | 8.9 GiB |
Default settings
The values the container ships with. They apply when you do not pass your own.
- Inference steps
- 25
- Guidance scale
- 4
Memory needed
The weights of Flex.1-alpha take 24.5 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 24.5 GiB (26,267,446,875 bytes).
$ sha256sum Flex.1-alpha.nbx
b8f01c26c73729a395c4c02f0db8e7c03690ecfbe0b94752d988c1de29bc172a
Questions
What is Flex.1-alpha?
A image generation model by ostris, distributed as a bfloat16 .nbx container and licensed Apache 2.0.
How big is the download?
24.5 GiB, of which 24.5 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 ostris/Flex.1-alpha, then neurobrix serve --model Flex.1-alpha.
CLI Usage
$ neurobrix import ostris/Flex.1-alpha
$ neurobrix serve --model Flex.1-alpha
$ neurobrix run --prompt "your prompt here"
$ neurobrix stop