feat(gallery): add FLUX.3 Image Klein and Klein Base (#12736)

Add the distilled and base FLUX.3 Image Klein models for the stablediffusion-ggml backend, each in Q4_K_M and Q8_0. They use the Qwen3-VL 4B text encoder and the FLUX.2 VAE. Klein runs with 16 steps and CFG 1.0, Klein Base with 20 steps and CFG 3.5, both with the euler sampler. The sd.cpp pin on master already contains FLUX.3 support. Assisted-by: Claude Code:claude-sonnet-5-5 Co-authored-by: Ettore Di Giacinto <mudler@localai.io>
2026-10-28 09:08:54 -04:00 · 2026-10-16 18:44:58 +02:00
1 parent eec2ec33de
commit d82624c81e
@@ -51667,6 +51667,196 @@
    - filename: stablediffusion-cpp/models/Qwen3-8B-Q4_K_M.gguf
      sha256: 120307ba529eb2439d6c430d94104dabd578497bc7bfe7e322b5d9933b449bd4
      uri: https://huggingface.co/unsloth/Qwen3-8B-GGUF/resolve/main/Qwen3-8B-Q4_K_M.gguf
+- name: flux.3-image-klein
+  url: github:mudler/LocalAI/gallery/flux-ggml.yaml@master
+  urls:
+    - https://huggingface.co/black-forest-labs/flux-3-image-klein
+    - https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs
+  description: |
+    FLUX.3 Image Klein is a compact FLUX.3 rectified flow transformer from Black Forest Labs that unifies text-to-image generation and image editing, with support for reference images. It uses Qwen3-VL 4B as the text encoder and the FLUX.2 VAE.
+
+    FLUX.3 Image Klein is the step-distilled variant, tuned for fast inference (default: 16 steps, CFG 1.0).
+
+    This entry uses a community Q4_K_M GGUF quantization of the diffusion model.
+  license: flux-kommunity-license-v1.0
+  icon: https://avatars.githubusercontent.com/u/164064024
+  tags:
+    - flux
+    - flux.3
+    - klein
+    - text-to-image
+    - image-generation
+    - image-editing
+    - gguf
+    - quantized
+    - diffusion
+    - black-forest-labs
+    - distilled
+  last_checked: "2026-10-16"
+  overrides:
+    cfg_scale: 1
+    options:
+      - diffusion_model
+      - vae_path:stablediffusion-cpp/models/flux2-vae.safetensors
+      - sampler:euler
+      - llm_path:stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      - offload_params_to_cpu:true
+      - vae_decode_only:false
+    parameters:
+      model: stablediffusion-cpp/models/flux3_image_klein_sd-Q4_K_M.gguf
+    step: 16
+  files:
+    - filename: stablediffusion-cpp/models/flux3_image_klein_sd-Q4_K_M.gguf
+      sha256: 2e429f54f9b2f2161b3123a615167ee6137aa21f1287f426dc895bb7647af5ec
+      uri: https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs/resolve/main/DISTILLED/flux3_image_klein_sd-Q4_K_M.gguf
+    - filename: stablediffusion-cpp/models/flux2-vae.safetensors
+      sha256: d64f3a68e1cc4f9f4e29b6e0da38a0204fe9a49f2d4053f0ec1fa1ca02f9c4b5
+      uri: https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensors
+    - filename: stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      sha256: 66358cb18bb6b3b1b6675aa412c7a88ef01d228f481184d13668e5201c730a0a
+      uri: https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/resolve/main/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+- name: flux.3-image-klein-q8_0
+  url: github:mudler/LocalAI/gallery/flux-ggml.yaml@master
+  urls:
+    - https://huggingface.co/black-forest-labs/flux-3-image-klein
+    - https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs
+  description: |
+    FLUX.3 Image Klein is a compact FLUX.3 rectified flow transformer from Black Forest Labs that unifies text-to-image generation and image editing, with support for reference images. It uses Qwen3-VL 4B as the text encoder and the FLUX.2 VAE.
+
+    FLUX.3 Image Klein is the step-distilled variant, tuned for fast inference (default: 16 steps, CFG 1.0).
+
+    This entry uses a community Q8_0 GGUF quantization of the diffusion model.
+  license: flux-kommunity-license-v1.0
+  icon: https://avatars.githubusercontent.com/u/164064024
+  tags:
+    - flux
+    - flux.3
+    - klein
+    - text-to-image
+    - image-generation
+    - image-editing
+    - gguf
+    - quantized
+    - diffusion
+    - black-forest-labs
+    - distilled
+  last_checked: "2026-10-16"
+  overrides:
+    cfg_scale: 1
+    options:
+      - diffusion_model
+      - vae_path:stablediffusion-cpp/models/flux2-vae.safetensors
+      - sampler:euler
+      - llm_path:stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      - offload_params_to_cpu:true
+      - vae_decode_only:false
+    parameters:
+      model: stablediffusion-cpp/models/flux3_image_klein_sd-Q8_0.gguf
+    step: 16
+  files:
+    - filename: stablediffusion-cpp/models/flux3_image_klein_sd-Q8_0.gguf
+      sha256: 216b01a2f8a6fcce947f70c2924e2e6a737da601c42e5a6e52a744e4a481b1ea
+      uri: https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs/resolve/main/DISTILLED/flux3_image_klein_sd-Q8_0.gguf
+    - filename: stablediffusion-cpp/models/flux2-vae.safetensors
+      sha256: d64f3a68e1cc4f9f4e29b6e0da38a0204fe9a49f2d4053f0ec1fa1ca02f9c4b5
+      uri: https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensors
+    - filename: stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      sha256: 66358cb18bb6b3b1b6675aa412c7a88ef01d228f481184d13668e5201c730a0a
+      uri: https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/resolve/main/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+- name: flux.3-image-klein-base
+  url: github:mudler/LocalAI/gallery/flux-ggml.yaml@master
+  urls:
+    - https://huggingface.co/black-forest-labs/flux-3-image-klein-base
+    - https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs
+  description: |
+    FLUX.3 Image Klein is a compact FLUX.3 rectified flow transformer from Black Forest Labs that unifies text-to-image generation and image editing, with support for reference images. It uses Qwen3-VL 4B as the text encoder and the FLUX.2 VAE.
+
+    FLUX.3 Image Klein Base is the undistilled variant. It runs with classifier-free guidance (default: 20 steps, CFG 3.5).
+
+    This entry uses a community Q4_K_M GGUF quantization of the diffusion model.
+  license: flux-kommunity-license-v1.0
+  icon: https://avatars.githubusercontent.com/u/164064024
+  tags:
+    - flux
+    - flux.3
+    - klein
+    - text-to-image
+    - image-generation
+    - image-editing
+    - gguf
+    - quantized
+    - diffusion
+    - black-forest-labs
+  last_checked: "2026-10-16"
+  overrides:
+    cfg_scale: 3.5
+    options:
+      - diffusion_model
+      - vae_path:stablediffusion-cpp/models/flux2-vae.safetensors
+      - sampler:euler
+      - llm_path:stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      - offload_params_to_cpu:true
+      - vae_decode_only:false
+    parameters:
+      model: stablediffusion-cpp/models/flux3_image_klein-Q4_K_M.gguf
+    step: 20
+  files:
+    - filename: stablediffusion-cpp/models/flux3_image_klein-Q4_K_M.gguf
+      sha256: b360892ffaee56ee4a4c76897d3b56977ca0183d233d57edccd9a1080200821b
+      uri: https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs/resolve/main/BASE/flux3_image_klein-Q4_K_M.gguf
+    - filename: stablediffusion-cpp/models/flux2-vae.safetensors
+      sha256: d64f3a68e1cc4f9f4e29b6e0da38a0204fe9a49f2d4053f0ec1fa1ca02f9c4b5
+      uri: https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensors
+    - filename: stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      sha256: 66358cb18bb6b3b1b6675aa412c7a88ef01d228f481184d13668e5201c730a0a
+      uri: https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/resolve/main/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+- name: flux.3-image-klein-base-q8_0
+  url: github:mudler/LocalAI/gallery/flux-ggml.yaml@master
+  urls:
+    - https://huggingface.co/black-forest-labs/flux-3-image-klein-base
+    - https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs
+  description: |
+    FLUX.3 Image Klein is a compact FLUX.3 rectified flow transformer from Black Forest Labs that unifies text-to-image generation and image editing, with support for reference images. It uses Qwen3-VL 4B as the text encoder and the FLUX.2 VAE.
+
+    FLUX.3 Image Klein Base is the undistilled variant. It runs with classifier-free guidance (default: 20 steps, CFG 3.5).
+
+    This entry uses a community Q8_0 GGUF quantization of the diffusion model.
+  license: flux-kommunity-license-v1.0
+  icon: https://avatars.githubusercontent.com/u/164064024
+  tags:
+    - flux
+    - flux.3
+    - klein
+    - text-to-image
+    - image-generation
+    - image-editing
+    - gguf
+    - quantized
+    - diffusion
+    - black-forest-labs
+  last_checked: "2026-10-16"
+  overrides:
+    cfg_scale: 3.5
+    options:
+      - diffusion_model
+      - vae_path:stablediffusion-cpp/models/flux2-vae.safetensors
+      - sampler:euler
+      - llm_path:stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      - offload_params_to_cpu:true
+      - vae_decode_only:false
+    parameters:
+      model: stablediffusion-cpp/models/flux3_image_klein-Q8_0.gguf
+    step: 20
+  files:
+    - filename: stablediffusion-cpp/models/flux3_image_klein-Q8_0.gguf
+      sha256: af0a5f00436c4e6911b7673b00803df0f7176c68b6ff4931a9629a51d5a3515d
+      uri: https://huggingface.co/realrebelai/FLUX-3-Image-Klein_GGUFs/resolve/main/BASE/flux3_image_klein-Q8_0.gguf
+    - filename: stablediffusion-cpp/models/flux2-vae.safetensors
+      sha256: d64f3a68e1cc4f9f4e29b6e0da38a0204fe9a49f2d4053f0ec1fa1ca02f9c4b5
+      uri: https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensors
+    - filename: stablediffusion-cpp/models/Qwen3VL-4B-Instruct-Q4_K_M.gguf
+      sha256: 66358cb18bb6b3b1b6675aa412c7a88ef01d228f481184d13668e5201c730a0a
+      uri: https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/resolve/main/Qwen3VL-4B-Instruct-Q4_K_M.gguf
 - name: Z-Image-Turbo
  url: github:mudler/LocalAI/gallery/z-image-ggml.yaml@master
  urls: