|
--- |
|
license: other |
|
license_name: flux-1-dev-non-commercial-license |
|
license_link: https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/LICENSE.md |
|
base_model: |
|
- black-forest-labs/FLUX.1-dev |
|
--- |
|
|
|
<img src="https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/resolve/main/images/banners/ComfyUI_2024-11-30_0088.jpg" alt="FLUX"> |
|
|
|
**====================================** |
|
|
|
**FLUX Model Merges & Tweaks: Detail Enhancement and Acceleration** |
|
|
|
<img src="https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/resolve/main/images/banners/ComfyUI_2024-11-30_0093.jpg" alt="FLUX Model Merges & Tweaks: Detail Enhancement and Acceleration"> |
|
|
|
This repository contains merged models, built upon the base models: |
|
- [Freepik's Flux.1-Lite-8B-alpha](https://huggingface.co/Freepik/flux.1-lite-8B-alpha) |
|
- [Flux-dev-de-distill](https://huggingface.co/nyanko7/flux-dev-de-distill) |
|
|
|
Detail enhancement and acceleration techniques have been applied, particularly optimized for NVIDIA 4XXX cards (maybe 3XXX too). The goal is to have high efficiency accelerated models with lower overhead. |
|
|
|
The (Detail Plus! De-Re-Distilled | Detail Plus! Distilled) models can be used with CFG remaining at 1-ish. Also the baked-in accelerators mostly work as intended. These are products of workflow optimizations, as such they can change/appear/disappear at any point and time. |
|
|
|
**====================================** |
|
|
|
<img src="https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/resolve/main/images/banners/ComfyUI_2024-11-30_0097.jpg" alt="Detail Plus!"> |
|
|
|
**Detail Plus! - Built upon the base model [Freepik's Flux.1-Lite-8B-alpha](https://huggingface.co/Freepik/flux.1-lite-8B-alpha):** |
|
|
|
**Detail Enhancement Used:** |
|
- **Style LORA - Extreme Detailer for FLUX.1-dev** (Weight: 0.5) ([Model Link](https://civitai.com/models/832683)) |
|
- **Best of Flux: Style Enhancing LoRA** (Weight: 0.25) ([Model Link](https://civitai.com/models/821668)) |
|
|
|
1. **SAFETensors Format (fp8_34m3fn_fast)**: |
|
- [flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-fp8_e4m3fn_fast](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-fp8_e4m3fn_fast.safetensors) |
|
- [flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-fp8_e4m3fn_fast](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-fp8_e4m3fn_fast.safetensors) |
|
- [flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-fp8_e4m3fn_fast](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-fp8_e4m3fn_fast.safetensors) |
|
|
|
3. **GGUF Quantized Models (Q8_0)**: |
|
- [flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus.Steps.Q8_0_quantized](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus.Steps.Q8_0_quantized.gguf) |
|
- [flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus.Steps.Q8_0_quantized](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus.Steps.Q8_0_quantized.gguf) |
|
- [flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus.Steps.Q8_0_quantized](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus.Steps.Q8_0_quantized.gguf) |
|
|
|
**====================================** |
|
|
|
**Detail Plus! Distilled - Built upon the base model [Freepik's Flux.1-Lite-8B-alpha](https://huggingface.co/Freepik/flux.1-lite-8B-alpha):** |
|
|
|
**Detail Enhancement Used:** |
|
- **Style LORA - Extreme Detailer for FLUX.1-dev** (Weight: 0.15) ([Model Link](https://civitai.com/models/832683)) |
|
- **Best of Flux: Style Enhancing LoRA** (Weight: 0.06) ([Model Link](https://civitai.com/models/821668)) |
|
|
|
**Distillation Used:** |
|
- **Flux distilled lora** (Weight: Hyper - 0.65 | Turbo - 0.50) ([Model Link](https://civitai.com/models/977247/flux-distilled-lora)) |
|
|
|
1. **SAFETensors Format (fp8_34m3fn_fast)**: |
|
- [flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/FP8/flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2.safetensors) |
|
- [flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/FP8/flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2.safetensors) |
|
- [flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/FP8/flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-Distilled-fp8_e4m3fn_fast-V2.safetensors) |
|
|
|
3. **GGUF Quantized Models (Q8_0)**: |
|
- [flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-Distilled-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/GGUF/Q8_0/flux.1-lite-8B-alpha-Hyper-8.Steps-Detail.Plus-Distilled-V2-Q8_0.gguf) |
|
- [flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-Distilled-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/GGUF/Q8_0/flux.1-lite-8B-alpha-Hyper-16.Steps-Detail.Plus-Distilled-V2-Q8_0.gguf) |
|
- [flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-Distilled-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/Lite-8B-Plus/Distilled/GGUF/Q8_0/flux.1-lite-8B-alpha-Turbo-8.Steps-Detail.Plus-Distilled-V2-Q8_0.gguf) |
|
|
|
**====================================** |
|
|
|
<img src="https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/resolve/main/images/banners/ComfyUI_2024-11-30_0119.jpg" alt="Detail Plus! De-Re-Distilled"> |
|
|
|
**Detail Plus! De-Re-Distilled (Built upon the base model [Flux-dev-de-distill](https://huggingface.co/nyanko7/flux-dev-de-distill):** |
|
|
|
**Detail Enhancement Used:** |
|
- **Style LORA - Extreme Detailer for FLUX.1-dev** (Weight: 0.15) ([Model Link](https://civitai.com/models/832683)) |
|
- **Best of Flux: Style Enhancing LoRA** (Weight: 0.15) ([Model Link](https://civitai.com/models/821668)) |
|
|
|
**Re-Distillation Used:** |
|
- **Flux distilled lora** (Weight: -1.00) ([Model Link](https://civitai.com/models/977247/flux-distilled-lora)) |
|
|
|
2. **SAFETensors Format V2 (fp8_34m3fn_fast)**: |
|
- [Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/FP8/Version%202/Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2.safetensors) |
|
- [Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/FP8/Version%202/Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2.safetensors) |
|
- [Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/FP8/Version%202/Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-fp8_e4m3fn_fast-V2.safetensors) |
|
|
|
5. **GGUF Quantized Models (Q8_0)**: |
|
- [Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q8_0/Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0.gguf) |
|
- [Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q8_0/Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0.gguf) |
|
- [Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q8_0/Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q8_0.gguf) |
|
|
|
6. **GGUF Quantized Models (Q6_K)**: |
|
- [Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q6_K/Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K.gguf) |
|
- [Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q6_K/Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K.gguf) |
|
- [Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q6_K/Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q6_K.gguf) |
|
|
|
7. **GGUF Quantized Models (Q4_K_S)**: |
|
- [Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q4_K_S/Flux-dev-de-re-distill-Hyper.8.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S.gguf) |
|
- [Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q4_K_S/Flux-dev-de-re-distill-Hyper.16.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S.gguf) |
|
- [Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/de-re-distill/GGUF/Q4_K_S/Flux-dev-de-re-distill-Turbo.8.Step-Detail.Plus-ReDis-MAIN-V2-Q4_K_S.gguf) |
|
|
|
**====================================** |
|
|
|
**Acceleration Credits:** |
|
|
|
* [ByteDance](https://huggingface.co/ByteDance), a leading AI technology company, optimized the following models: |
|
- [Hyper-FLUX.1-dev-16steps](https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-FLUX.1-dev-16steps-lora.safetensors) |
|
- [Hyper-FLUX.1-dev-8steps](https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-FLUX.1-dev-8steps-lora.safetensors) |
|
|
|
* [Alimama Creative](https://huggingface.co/alimama-creative), a renowned NLP innovator, optimized the following model: |
|
- [FLUX.1-Turbo-Alpha](https://huggingface.co/alimama-creative/FLUX.1-Turbo-Alpha) |
|
|
|
**====================================** |
|
|
|
**Attribution and Licensing Notice:** |
|
|
|
The [FLUX.1-dev Model](https://huggingface.co/black-forest-labs/FLUX.1-dev) is licensed by Black Forest Labs, Inc. under the FLUX.1-dev [Non-Commercial License](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md). Copyright Black Forest Labs, Inc. |
|
|
|
Our model weights are released under the FLUX.1-dev [Non-Commercial License](https://huggingface.co/mdmachine/FLUX.Model.Merge-Detail.Enhancement.and.Acceleration/blob/main/LICENSE.md). |
|
|
|
This merge combines the strengths of these models, applying detail enhancement and acceleration techniques to create a unique and powerful AI model built upon [Freepik's Flux.1-Lite-8B-alpha](https://huggingface.co/Freepik/flux.1-lite-8B-alpha) & [Flux-dev-de-distill](https://huggingface.co/nyanko7/flux-dev-de-distill). We hope this contributes positively to the development of the NLP community! |