Model Gallery

9 models from 1 repositories

Filter by type:

Filter by tags:

qwen-image
We are thrilled to release Qwen-Image, an image generation foundation model in the Qwen series that achieves significant advances in complex text rendering and precise image editing. Experiments show strong general capabilities in both image generation and editing, with exceptional performance in text rendering, especially for Chinese.

Repository: localaiLicense: apache-2.0

qwen-image-edit
Qwen-Image-Edit is a model for image editing, which is based on Qwen-Image.

Repository: localaiLicense: apache-2.0

qwen-image-edit-2509
Qwen-Image-Edit is a model for image editing, which is based on Qwen-Image.

Repository: localaiLicense: apache-2.0

qwen-image-2.1-q4_k-ggml
Qwen-Image 2.1 is the Qwen image generation foundation model, with strong prompt adherence and accurate text rendering (English and Chinese). It uses Qwen3-VL-8B as the text encoder and its own VAE, and supports both text-to-image and image editing: pass reference images to edit them. This is the Q4_K (4-bit) quantization (~4.2GB diffusion model) by leejet for stable-diffusion.cpp. The bundle also pulls the Qwen3-VL-8B-Instruct text encoder with its vision projector (used for image editing) and the Qwen-Image 2.1 VAE. Use image dimensions divisible by 32.

Repository: localaiLicense: qwen-research

qwen-image-2.1-q8_0-ggml
Qwen-Image 2.1 is the Qwen image generation foundation model, with strong prompt adherence and accurate text rendering (English and Chinese). It uses Qwen3-VL-8B as the text encoder and its own VAE, and supports both text-to-image and image editing: pass reference images to edit them. This is the Q8_0 (8-bit) quantization (~7.7GB diffusion model) by leejet for stable-diffusion.cpp. The bundle also pulls the Qwen3-VL-8B-Instruct text encoder with its vision projector (used for image editing) and the Qwen-Image 2.1 VAE. Use image dimensions divisible by 32.

Repository: localaiLicense: qwen-research

qwen-image-2.1-anime-fusal-gguf
Qwen-Image 2.1 Anime FUSAL runtime assembly with a Q5_0 denoiser, Heretic Q4_K_M text encoder, BF16 VAE, and separate rank-80 FUSAL adapter. Downloads total about 10.4 GiB; allow at least that much combined RAM/VRAM plus runtime overhead. This assembly has mixed licensing: the Qwen-derived denoiser, VAE, and FUSAL adapter are under the Qwen Research License and are limited to non-commercial research or evaluation unless separately authorized; the Heretic encoder is Apache-2.0. Review the repository notices for NSFW content and third-party-rights responsibilities. Use 30 steps, Euler sampling, and CFG 1.0. The adapter is applied at runtime, so prompts must use `fusal style. [TASK: TEXT_TO_IMAGE] DESCRIPTION`, replacing DESCRIPTION with the image description. This practical encoder omits vision weights, so this entry is text-to-image only and does not support reference-image editing.

Repository: localaiLicense: other

qwen-image-2.1-turbo-uncensored-gguf
Qwen-Image 2.1 Turbo text-to-image assembly with AtomicChat's AD-Q5_K denoiser, Q8_0 abliterated text encoder, and the Qwen Image 2.1 BF16 VAE. It uses the Turbo runtime's exact eight-step sigma schedule with Euler sampling and CFG 1.0. The denoiser and VAE are governed by the Qwen Research License; the separately modified encoder is Apache-2.0. The complete image runtime is therefore not Apache-2.0. AtomicChat reports that changing the encoder's refusal direction produced no measured category-outcome change in its image tests, so this entry does not promise broader image capability. This bundle has no vision projector and supports text-to-image only, not reference-image editing.

Repository: localaiLicense: other

qwen-image-2.1-uncensored
🤖 ModelScope  |   🤗 HuggingFace  |   📑 Blog  |   🖥️ Demo  |   🫨 Discord  |   💬 WeChat ## Introduction We are excited to open-source **Qwen-Image-2.1**, a unified text-to-image generation and image editing model in the Qwen family. With just **7B parameters in its visual generation component** (32 Single-Stream DiT layers), Qwen-Image-2.1 balances generation quality, inference efficiency, and versatility. Four key improvements define this release: ...

Repository: localaiLicense: other

qwen-image-2.1-text-encoder-heretic
> 🍱 **All-GGUF Qwen-Image-2.1:** text encoder GGUF + DiT GGUF (Q8_0 / Q6_K / Q4_K_M, loads in stock ComfyUI-GGUF) + PE-T2I GGUF. > [!IMPORTANT] > **The 4 `model-0000x-of-00004` shards are the HF `transformers` checkpoint (bf16) — they do not load in ComfyUI.** > ComfyUI text encoders use a different key layout (the `model.language_model.` prefix is dropped > when repacking), so pointing ComfyUI at these shards will not work. > > For ComfyUI use one of these instead — all keep the vision tower, which 2.1 needs for editing. > **Mac / non-CUDA:** use the single file `qwen3vl_8b_bf16_heretic.safetensors` in *this* repo (full bf16, 17.5 GB — also in the GGUF repo) — > put it in `models/text_encoders/` and load it with the stock `CLIPLoader`, type `qwen_image`, then `TextEncodeQwenImage21`. > It is the only format here that needs no CUDA-specific kernels. > > | Repo | File | Loader | > |---|---|---| > | **this repo** | `qwen3vl_8b_bf16_heretic.safetensors` (bf16 single file, any device) | `CLIPLoader`, type `qwen_image` | > | `…-NVFP4` | `qwen3vl_8b_nvfp4_heretic.safetensors` | `CLIPLoader`, type `qwen_image` | > | `…-W4A8` | `qwen3vl_8b_w4a8_heretic.safetensors` | `CLIPL ...

Repository: localaiLicense: apache-2.0