license: apache-2.0
base_model: ByteDance-Seed/SeedVR2-3B
pipeline_tag: image-to-image
library_name: mlx-gen
tags:
- mlx
- mlx-gen
- mflux
- apple-silicon
- 8-bit
- seedvr2
- image-upscaling
- super-resolution
seedvr2-3b-8bit
This repository contains MLX-Gen saved weights for ByteDance-Seed/SeedVR2-3B. The checkpoint is designed for local Apple Silicon inference with mlx-gen.
It uses the mflux/MLX saved-weight layout. Quantized checkpoints include MLX quantization tensors. It is not a Diffusers or Transformers from_pretrained() checkpoint.
Source Model
Original model: ByteDance-Seed/SeedVR2-3B.
License and Access
This quantized derivative follows the Apache 2.0 license of the source model.
Quantization
This is an MLX 8-bit checkpoint for SeedVR2 3B image super-resolution.
- q8 for quantizable SeedVR2 transformer linears and VAE attention linears.
- BF16 for convolutions, normalization layers, and other non-quantizable parameters.
- The static SeedVR2 positive embedding is provided by the MLX-Gen runtime.
See the MLX-Gen quantization docs for compatibility notes and measured behavior.
Validation
Measured on an Apple M5 Max with 128 GB unified memory using mlx-gen 0.18.12.
The profile upscales docs/assets/upscaling/seedvr2-5x-source.jpg from 133x113
to 658x560 with --resolution 5x --seed 42 --metadata.
| Package | Storage | Generation time | Wall time | Max RSS |
|---|---|---|---|---|
ByteDance-Seed/SeedVR2-3B source generation files |
13.57 GiB | 2.27 s | 5.89 s | 25.49 GiB |
AbstractFramework/seedvr2-3b-8bit |
4.39 GiB | 2.13 s | 3.22 s | 4.73 GiB |
Compatibility
Requires mlx-gen >= 0.18.12.
Generated with mlx-gen 0.18.12.
Use mlxgen download and mflux-upscale-seedvr2 for this SeedVR2 package.
Usage
python -m pip install -U mlx-gen
mlxgen download --model AbstractFramework/seedvr2-3b-8bit
mflux-upscale-seedvr2 \
--model AbstractFramework/seedvr2-3b-8bit \
--image-path input.png \
--resolution 2x \
--seed 42 \
--metadata \
--output upscaled.png
Attribution
MLX-Gen is based on mflux by Filip Strand and the original mflux contributors. This model card is generated by MLX-Gen so derived checkpoints keep that attribution visible.
Quantized and contributed by @lpalbou.