Buckets:
| # FLUX.2 & Klein 9B RefMods Documentation | |
| Welcome to the official documentation and reference guides for **FLUX.2 & Klein 9B RefMods (Reference Latent Adapters)**. | |
| - **Interactive Model Browser:** [https://huggingface.co/spaces/malcolmrey/browser](https://huggingface.co/spaces/malcolmrey/browser) (Visual search & previews across 1,400+ models) | |
| - **Model Weights Hub:** [https://huggingface.co/malcolmrey/klein9](https://huggingface.co/malcolmrey/klein9) | |
| - **ComfyUI Custom Node:** [https://github.com/malcolmamal/ComfyUI-Flux2Klein9Mod](https://github.com/malcolmamal/ComfyUI-Flux2Klein9Mod) | |
| - **Ready-to-Use Workflows:** [`../workflows/`](../workflows/) (`workflow_klein9_refmod.json` & `workflow_klein9_refmod_lora.json`) | |
| --- | |
| ## 📚 Guides & Technical Documentation | |
| 1. **[FLUX.2 & Klein 9B RefMods: Architecture, Extraction, and Usage Guide](FLUX2_KLEIN9_REFMODS_GUIDE.md)** | |
| The primary comprehensive architectural handbook covering: | |
| - Theoretical underpinnings of 2D RoPE reference latent injection vs legacy pixel encoding | |
| - Where to find and download 1,400+ pre-extracted RefMods on Hugging Face ([malcolmrey/klein9](https://huggingface.co/malcolmrey/klein9) & [malcolmrey/browser](https://huggingface.co/spaces/malcolmrey/browser)) | |
| - Installing and configuring the `ComfyUI-Flux2Klein9Mod` custom node | |
| - Detailed breakdown of both official ComfyUI JSON workflows (`workflow_klein9_refmod.json` & `workflow_klein9_refmod_lora.json`) | |
| - Step-by-step RefMod creation with dataset curation best practices and CLI extraction | |
| 2. **[Visual Benchmarks & 5-Column Comparison Studies](VISUAL_BENCHMARKS_AND_COMPARISONS.md)** | |
| Detailed empirical evaluations across 6 celebrity benchmark datasets: | |
| - Embedded 5-column comparison grids for all benchmark subjects (Brett Cooper, Alison Brie, Olivia Wilde, Jodie Foster, Rebecca Ferguson, Sydney Sweeney) | |
| - Direct download links to individual model weights (`.safetensors`) on Hugging Face | |
| - High-resolution single run artifact links ([`examples/`](examples/)) for Pure LoRA, 8-Raw Reference, Pure RefMod, and Hybrid stacks | |
| - Computational throughput, token load, and VRAM scaling breakdowns | |
| --- | |
| ## 🎛️ Official ComfyUI Workflows (`workflows/`) | |
| Drag-and-drop workflow files ready for immediate use in ComfyUI: | |
| - **[`workflow_klein9_refmod.json`](../workflows/workflow_klein9_refmod.json):** Pure RefMod Text-to-Image generation for FLUX.2 / Klein 9B. | |
| - **[`workflow_klein9_refmod_lora.json`](../workflows/workflow_klein9_refmod_lora.json):** Hybrid RefMod + LoRA Power Stack for maximum facial likeness and micro-texture fidelity. | |
| --- | |
| ## 🛠️ Tools & Scripts | |
| - **Generator & Batch Extractor Script:** [`../generate_flux2_klein9_refmod.py`](../generate_flux2_klein9_refmod.py) | |
| CLI generator tool supporting single-dataset extraction, custom token budgets, dry-run inspection, and high-speed mass batch extraction. | |
| - **ComfyUI Custom Node:** [`../comfyui/custom_nodes/ComfyUI-Flux2Klein9Mod/`](../comfyui/custom_nodes/ComfyUI-Flux2Klein9Mod/) | |
| GitHub Repository: [github.com/malcolmamal/ComfyUI-Flux2Klein9Mod](https://github.com/malcolmamal/ComfyUI-Flux2Klein9Mod) | |
| - **Hugging Face Model Hub:** [huggingface.co/malcolmrey/klein9](https://huggingface.co/malcolmrey/klein9) | |
| 1,400+ open-weight RefMods and Prodigy LoRAs. | |
Xet Storage Details
- Size:
- 3.33 kB
- Xet hash:
- 058c9737689ca329dded29fcbda68422b34f042ba830dbbb4b59d69cc7450bbe
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.