File size: 11,631 Bytes
4966da1 42217f6 530b61a 4966da1 a8ca6b8 a41185c 276c295 a4b8307 9ccdc7c 271bb72 4966da1 a41185c 9489fe0 271bb72 9489fe0 009763c 271bb72 a4b8307 276c295 a41185c 276c295 271bb72 9489fe0 276c295 42217f6 276c295 42217f6 4966da1 a4b8307 276c295 9235f2f 009763c 9235f2f 12586d7 9235f2f a41185c 4966da1 e74fe76 271bb72 d34432f 271bb72 276c295 009763c 9235f2f 276c295 009763c 276c295 009763c c34d875 009763c a4b8307 276c295 a4b8307 4966da1 d34432f 4966da1 a8ca6b8 4966da1 a8ca6b8 4966da1 a4b8307 276c295 4966da1 276c295 4966da1 276c295 4966da1 276c295 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 |
---
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 (de-re-destill) model can be used with CFG remaining at 1. Also the baked-in accelerators work as intended.
**====================================**
<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! |