metadata
name: Mystic-MLX-Q8
base_model: black-forest-labs/FLUX.1-dev
license: other
pipeline_tag: text-to-image
tasks:
- text-to-image
- image-to-image
- image-generation
language: en
get_started_code: >-
uvx --from mflux mflux-generate --model exdysa/Mystic-MLX-Q8 --prompt 'Test
Prompt' --base-model dev --steps 16 --guidance 7.5 --seed 10 --width 1024
--height 1024 -q 8 --prompt 'Test prompt'
library_name: mlx
tags:
- flux-1.dev
- flux
- mlx
- apple
- mystic
Original Model Link : https://huggingface.co/enhanceaiteam/Mystic
name: Mystic-MLX-Q8
base_model: black-forest-labs/FLUX.1-dev
license: other
pipeline_tag: text-to-image
tasks :
- text-to-image
- image-to-image
- image-generation
language: en
get_started_code: uvx --from mflux mflux-generate --model exdysa/Mystic-MLX-Q8 --prompt 'Test Prompt' --base-model dev --steps 16 --guidance 7.5 --seed 10 --width 1024 --height 1024 -q 8 --prompt 'Test prompt'
library_name: mlx
tags:
- flux-1.dev
- flux
- mlx
- apple
- mystic
Mystic-MLX-Q8
Mystic-MLX-Q8 is a Flux.1-dev fine-tune that lowers steps required for convergence while maintaining quality outputs. This repo contains the MLX converted model weights.
MLX is a framework for METAL graphics supported by Apple computers with ARM M-series processors (M1/M2/M3/M4)
Generation using uv https://docs.astral.sh/uv/**:
uvx --from mflux mflux-generate --model exdysa/Mystic-MLX-Q8 --prompt 'Test Prompt' --base-model dev --steps 16 --guidance 7.5 --seed 10 --width 1024 --height 1024 -q 8
Generation using pip:
pipx --from mflux mflux-generate --model exdysa/Mystic-MLX-Q8 --prompt 'Test Prompt' --base-model dev --steps 16 --guidance 7.5 --seed 10 --width 1024 --height 1024 -q 8