README.md
7.1 KB · 122 lines · markdown Raw
1 ---
2 license: other
3 license_name: qwen-research
4 base_model:
5 - Qwen/Qwen-Image-2.1
6 base_model_relation: quantized
7 pipeline_tag: text-to-image
8 library_name: gguf
9 tags:
10 - gguf
11 - qwen
12 - image-generation
13 - comfyui
14 - comfyui-gguf
15 ---
16
17 > [!IMPORTANT]
18 > The uncensored GGUFs are now available!
19 > If you encounter any issues or would like to share your generated images, feel free to post them in the Community section.
20
21 # Qwen-Image-2.1 Uncensored GGUF
22
23 GGUF quantizations of [Qwen/Qwen-Image-2.1](https://huggingface.co/Qwen/Qwen-Image-2.1) for local image generation using the original upstream base weights.
24
25 ## Benchmark
26
27 ![Qwen-Image-2.1 benchmark](assets/Qwen-Image-2.1-Benchmark.png)
28
29 ## Uncensored GGUF Files
30
31 | **Quantization** | **File** | **Size** |
32 | --- | --- | ---: |
33 | BF16 | [**qwen-image-2.1-UC-BF16.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-BF16.gguf) | 14.23 GB |
34 | FP8 | [**qwen-image-2.1-UC-fp8.safetensors**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-fp8.safetensors) | 6.63 GB |
35 | INT8 ConvRot | [**qwen-image-2.1-UC-int8_convrot.safetensors**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-int8_convrot.safetensors) | 6.76 GB |
36 | Q8_0 | [**qwen-image-2.1-UC-Q8_0.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-Q8_0.gguf) | 7.59 GB |
37 | Q6_K | [**qwen-image-2.1-UC-Q6_K.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-Q6_K.gguf) | 5.88 GB |
38 | Q5_K_M | [**qwen-image-2.1-UC-Q5_K_M.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-Q5_K_M.gguf) | 5.22 GB |
39 | Q4_K_M | [**qwen-image-2.1-UC-Q4_K_M.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-Q4_K_M.gguf) | 4.60 GB |
40 | Q4_0 | [**qwen-image-2.1-UC-Q4_0.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/qwen-image-2.1-UC-Q4_0.gguf) | 4.15 GB |
41
42 **Q4_K_M** is recommended for the best balance of size and quality.
43
44 ## GGUF files
45
46 | **Quantization** | **File** | **Size** |
47 | --- | --- | ---: |
48 | Q8_0 | [**qwen-image-2.1-Q8_0.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/base/qwen-image-2.1-Q8_0.gguf) | 7.59 GB |
49 | Q6_K | [**qwen-image-2.1-Q6_K.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/base/qwen-image-2.1-Q6_K.gguf) | 5.88 GB |
50 | Q5_K_M | [**qwen-image-2.1-Q5_K_M.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/base/qwen-image-2.1-Q5_K_M.gguf) | 5.22 GB |
51 | Q4_K_M | [**qwen-image-2.1-Q4_K_M.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/base/qwen-image-2.1-Q4_K_M.gguf) | 4.60 GB |
52 | Q4_0 | [**qwen-image-2.1-Q4_0.gguf**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/base/qwen-image-2.1-Q4_0.gguf) | 4.05 GB |
53
54
55 ## Text Encoders & VAE
56
57 Companion model files packaged for ComfyUI:
58
59 | **Type** | **File** | **Precision** | **Size** |
60 | --- | --- | --- | ---: |
61 | Text Encoder | [**text_encoders/qwen3vl_8b_bf16.safetensors**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/text_encoders/qwen3vl_8b_bf16.safetensors) | BF16 | 17.53 GB |
62 | Text Encoder | [**text_encoders/qwen3vl_8b_int8_convrot.safetensors**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/text_encoders/qwen3vl_8b_int8_convrot.safetensors) | Int8 | 9.35 GB |
63 | VAE | [**vae/qwen_image_2.1_vae_bf16.safetensors**](https://huggingface.co/abenzerps/Qwen-Image-2.1-Uncensored-GGUF/blob/main/vae/qwen_image_2.1_vae_bf16.safetensors) | BF16 | 676 MB |
64
65 ## Usage
66
67 Use the model with [ComfyUI](https://github.com/comfyanonymous/ComfyUI) and [ComfyUI-GGUF](https://github.com/leejet/ComfyUI-GGUF).
68
69 All required companion files (GGUF transformer, text encoder, and VAE) are hosted directly in this repository.
70
71 ### 1. Download & File Placement
72
73 Download the files and place them in their respective ComfyUI directories:
74
75 ```text
76 ComfyUI/
77 └── models/
78 ├── diffusion_models/
79 │ └── qwen-image-2.1-UC-Q4_K_M.gguf # Choose one GGUF quantization (Q4_K_M recommended)
80 ├── text_encoders/
81 │ └── qwen3vl_8b_bf16.safetensors # Or qwen3vl_8b_int8_convrot.safetensors (recommended for lower memory)
82 └── vae/
83 └── qwen_image_2.1_vae_bf16.safetensors
84 ```
85
86 ### 2. ComfyUI Setup
87
88 1. **Install ComfyUI-GGUF**: Use the maintained fork with native Qwen-Image 2.1 support by cloning [leejet/ComfyUI-GGUF](https://github.com/leejet/ComfyUI-GGUF) into your custom nodes:
89 ```bash
90 cd ComfyUI/custom_nodes
91 git clone https://github.com/leejet/ComfyUI-GGUF
92 ```
93 *(Note: If you have the older `city96/ComfyUI-GGUF` installed and encounter an `Unknown model architecture!` error, update to the `leejet` fork above or add `ModelQwenImage` to `tools/convert.py`).*
94 2. **Node Configuration**:
95 - **Diffusion Model**: Add the **`Unet Loader (GGUF)`** node and select your downloaded `.gguf` file.
96 - **Text Encoder**: Add the standard **`CLIPLoader`** node, select `qwen3vl_8b_bf16.safetensors` (or `int8`), and set **`type`** to **`qwen_image`**.
97 - **VAE**: Add the standard **`VAELoader`** node and select `qwen_image_2.1_vae_bf16.safetensors`.
98 3. **Official Workflows**:
99 - You can use the official Comfy-Org workflow templates: [Text-to-Image](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/image_qwen_image_2_1_t2i.json) or [Image Edit](https://github.com/Comfy-Org/workflow_templates/blob/main/templates/image_qwen_image_2_1_image_edit.json).
100 - In the workflow, simply replace the default `UNETLoader` node with **`Unet Loader (GGUF)`**.
101
102 ### Memory & Performance Notes
103
104 - **Optimal Setup (GPU + RAM)**: Keep the **GGUF diffusion model in GPU VRAM** (where speed is crucial during sampling) and let the **text encoder run in / offload to System RAM (CPU)**. Because text encoding only runs once per prompt, this saves 9–17 GB of VRAM with virtually zero impact on generation speed.
105 - **Recommended Configuration**:
106 - **Diffusion**: `qwen-image-2.1-UC-Q4_K_M.gguf` (~4.6 GB in VRAM)
107 - **Text Encoder**: `qwen3vl_8b_int8_convrot.safetensors` (~9.35 GB in RAM)
108 - **Low VRAM Mode**: If you experience VRAM out-of-memory errors, start ComfyUI with the `--lowvram` argument.
109
110 ## Uncensored
111
112 This GGUF release has no built-in safety checker or content filter. It generates adult, NSFW, and sensitive imagery directly without prompt refusals or blacked-out images. Output behavior depends solely on the input prompts and the environment in which the model is executed.
113
114 ## Source and build
115
116 - **Source model:** [Qwen/Qwen-Image-2.1](https://huggingface.co/Qwen/Qwen-Image-2.1)
117 - **Text encoder & VAE source:** [Comfy-Org/Qwen-Image-2.1](https://huggingface.co/Comfy-Org/Qwen-Image-2.1)
118 - **Source revision:** `b3179ad355be050328e483a9dfdd9e60cd62adfa`
119 - **Conversion:** [stable-diffusion.cpp](https://github.com/leejet/stable-diffusion.cpp) commit `1330cebae8f2ba99249df846cc0c9444fcbd4308`
120 - **License:** Qwen Research License
121 - **Checksums:** [SHA256SUMS](./SHA256SUMS)
122