mflux

v.0.18.0 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

Published 1mo Model Serving & MLOps

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai apple-silicon diffusers fibo flux huggingface

+7 more

machine-learning mlx qwen qwen-image seedvr2 transformers z-image

Summary

AI summary

Updates 👩‍💻 Contributors, 🐛 Bug Fixes, and 🎨 New Model Support across a mixed release.

Changes in this release

Type	Severity	Summary	CVE
Feature	Low	Adds ERNIE-Image and ERNIE-Image-Turbo model support with CLI entrypoints, LoRA inference/training, and mflux-save. Adds ERNIE-Image and ERNIE-Image-Turbo model support with CLI entrypoints, LoRA inference/training, and mflux-save. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Low	Adds Ideogram 4 FP8 text-to-image support with JSON caption handling and FP8 safetensors loading. Adds Ideogram 4 FP8 text-to-image support with JSON caption handling and FP8 safetensors loading. Source: llm_adapter@2026-06-07 Confidence: high	—
Performance	Medium	Improves FLUX.2 Klein 9B KV-cache, achieving ~2.4× speedup on multi-reference edit workloads. Improves FLUX.2 Klein 9B KV-cache, achieving ~2.4× speedup on multi-reference edit workloads. Source: llm_adapter@2026-06-07 Confidence: high	—
Bugfix
Bugfix	Medium	Allows --guidance > 1.0 for FLUX.2 Klein edits by checking resolved model config instead of requiring base model name. Allows --guidance > 1.0 for FLUX.2 Klein edits by checking resolved model config instead of requiring base model name. Source: llm_adapter@2026-06-07 Confidence: high	—
Bugfix	Medium	Fixes FileNotFoundError: text_encoder_2 in mflux-generate by routing Klein models through Flux2Klein and skipping unused T5/text_encoder_2 path. Fixes FileNotFoundError: text_encoder_2 in mflux-generate by routing Klein models through Flux2Klein and skipping unused T5/text_encoder_2 path. Source: llm_adapter@2026-06-07 Confidence: high	—
Bugfix	Medium	Evicts text encoder after encoding and clears MLX cache between seeds to prevent OOM on large models like FLUX.2 Klein 9B. Evicts text encoder after encoding and clears MLX cache between seeds to prevent OOM on large models like FLUX.2 Klein 9B. Source: llm_adapter@2026-06-07 Confidence: high	—

Full changelog

🎨 New Model Support

ERNIE-Image & ERNIE-Image-Turbo: Port Baidu's ERNIE-Image models with text-to-image CLI entrypoints, LoRA inference and training, and mflux-save support.
Ideogram 4 FP8: Add Ideogram 4 FP8 text-to-image support with JSON caption handling, FP8 safetensors loading, and a dedicated CLI entrypoint.

✨ Improvements

FLUX.2 Klein 9B KV-cache: Add KV-cache support for flux2-klein-9b-kv with roughly 2.4× speedup on multi-reference edit workloads.

🐛 Bug Fixes

FLUX.2 Klein Edit guidance: Allow --guidance > 1.0 for FLUX.2 Klein edits by checking the resolved FLUX.2 model config instead of requiring a base model name; defaults remain unchanged.
FLUX.2 Klein mflux-generate: Fix FileNotFoundError: text_encoder_2 by routing Klein models through Flux2Klein and skipping the unused T5/text_encoder_2 weight path.
Memory management: Evict the text encoder after encoding and clear the MLX cache between seeds on multi-seed runs to prevent OOM on large models such as FLUX.2 Klein 9B.

📝 Documentation

Related projects: Add mlx-taef and mlx-teacache to the Related projects list.

👩‍💻 Contributors

@azrahello
@c2p-cmd
@IonDen
@lpalbou
@michaeltrefry
@omercelik
@plz12345

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track mflux

Get notified when new releases ship.

About mflux

MLX native implementations of state-of-the-art generative image models

All releases →