This model requires you to accept the Apache 2.0 license terms before download.
Description
SwinIR Classical x4 - Swin Transformer image-restoration model for 4x super-resolution, reconstructing sharp high-resolution detail from low-resolution images. Works at any input size. Apache 2.0.
Tags
At a glance
| Kind | image upscaler |
|---|---|
| Precision | float32 |
| Container size | 43.2 MiB |
| Weights inside | 56.7 MiB |
| Components | 1 |
| Licence | Apache 2.0 |
| Origin | JingyunLiang/SwinIR @ unrecorded |
What is inside the container
A .nbx container packs every part of SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4 — its single component with weights, graphs and runtime settings — so one file is all the runtime needs.
| Component | Blocks | Layers | Weights |
|---|---|---|---|
| model | 0 | — | 56.7 MiB |
Memory needed
The weights of SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4 take 56.7 MiB in float32. 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 43.2 MiB (45,347,919 bytes).
$ sha256sum SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4.nbx
20a764fa500984a383f95a7fba9f029ee3a5d5fccb9f16b2bfa52c29a5416e86
Questions
What is SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4?
A image upscaler by JingyunLiang, distributed as a float32 .nbx container and licensed Apache 2.0.
How big is the download?
43.2 MiB, of which 56.7 MiB are weights.
How do I run it?
Install the runtime from the download page, then use the commands under “CLI Usage” below: neurobrix import JingyunLiang/SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4, then neurobrix serve --model SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4.
CLI Usage
$ neurobrix import JingyunLiang/SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4
$ neurobrix serve --model SwinIR-001_classicalSR_DIV2K_s48w8_SwinIR-M_x4
$ neurobrix run --prompt "your prompt here"
$ neurobrix stop