Add SeFi-Image pipeline

docs/source/en/_toctree.yml

@@ -389,6 +389,8 @@
         title: SanaVideoTransformer3DModel
       - local: api/models/sd3_transformer2d
         title: SD3Transformer2DModel
+      - local: api/models/sefi_transformer2d
+        title: SeFiTransformer2DModel
       - local: api/models/skyreels_v2_transformer_3d
         title: SkyReelsV2Transformer3DModel
       - local: api/models/stable_audio_transformer
@@ -605,6 +607,8 @@
         title: Sana Sprint
       - local: api/pipelines/sana_video
         title: Sana Video
+      - local: api/pipelines/sefi
+        title: SeFi-Image
       - local: api/pipelines/shap_e
         title: Shap-E
       - local: api/pipelines/stable_cascade

docs/source/en/api/models/sefi_transformer2d.md

@@ -0,0 +1,21 @@
+<!--Copyright 2026 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# SeFiTransformer2DModel
+
+[`SeFiTransformer2DModel`] is the SeFi-Image transformer. It uses a Flux2-style backbone with a dual timestep embedding
+for semantic and texture latent streams.
+
+## SeFiTransformer2DModel
+
+[[autodoc]] SeFiTransformer2DModel
+  - forward

docs/source/en/api/pipelines/sefi.md

@@ -0,0 +1,52 @@
+<!--Copyright 2026 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# SeFi-Image
+
+[SeFi-Image](https://github.com/jmliu206/SeFi-Image) is a text-to-image model family based on Semantic-First Diffusion.
+It denoises a semantic latent stream slightly ahead of a texture latent stream, then decodes the final texture latents
+to images.
+
+The public checkpoints on the Hub are gated and distributed under a non-commercial license. Accept the checkpoint
+license on the Hub before downloading them. Original SeFi-Image artifacts can be converted with
+`scripts/convert_sefi_to_diffusers.py` before loading them with [`SeFiPipeline`].
+
+```python
+import torch
+from diffusers import SeFiPipeline
+
+pipe = SeFiPipeline.from_pretrained("./sefi-1b-base-diffusers", torch_dtype=torch.bfloat16)
+pipe.to("cuda")
+
+image = pipe("A red apple on a wooden table.").images[0]
+image.save("sefi.png")
+```
+
+Convert the original artifacts first:
+
+```bash
+python scripts/convert_sefi_to_diffusers.py \
+  --checkpoint SeFi-Image/SeFi-Image-1B-Base \
+  --output ./sefi-1b-base-diffusers \
+  --token $HF_TOKEN
+```
+
+Turbo checkpoints are distilled and support 4, 8, or 10 inference steps with `guidance_scale=1.0`.
+
+## SeFiPipeline
+
+[[autodoc]] SeFiPipeline
+  - all
+
+## SeFiPipelineOutput
+
+[[autodoc]] pipelines.sefi.pipeline_output.SeFiPipelineOutput