add a note on how to run pipeline without the componen4ts manager

* up

* style

docs/source/en/modular_diffusers/modular_diffusers_states.md

@@ -25,9 +25,7 @@ This guide explains how states work and how they connect blocks.
 
 The [`~modular_pipelines.PipelineState`] is a global state container for all blocks. It maintains the complete runtime state of the pipeline and provides a structured way for blocks to read from and write to shared data.
 
-There are two dict's in [`~modular_pipelines.PipelineState`] for structuring data.
-
-- The `values` dict is a **mutable** state containing a copy of user provided input values and intermediate output values generated by blocks. If a block modifies an `input`, it will be reflected in the `values` dict after calling `set_block_state`.
+[`~modular_pipelines.PipelineState`] stores all data in a `values` dict, which is a **mutable** state containing user provided input values and intermediate output values generated by blocks. If a block modifies an `input`, it will be reflected in the `values` dict after calling `set_block_state`.
 
 ```py
 PipelineState(

docs/source/en/modular_diffusers/modular_pipeline.md

@@ -12,27 +12,28 @@ specific language governing permissions and limitations under the License.
 
 # ModularPipeline
 
-[`ModularPipeline`] converts [`~modular_pipelines.ModularPipelineBlocks`]'s into an executable pipeline that loads models and performs the computation steps defined in the block. It is the main interface for running a pipeline and it is very similar to the [`DiffusionPipeline`] API.
+[`ModularPipeline`] converts [`~modular_pipelines.ModularPipelineBlocks`] into an executable pipeline that loads models and performs the computation steps defined in the blocks. It is the main interface for running a pipeline and the API is very similar to [`DiffusionPipeline`] but with a few key differences.
 
-The main difference is to include an expected `output` argument in the pipeline.
+- **Loading is lazy.** With [`DiffusionPipeline`], [`~DiffusionPipeline.from_pretrained`] creates the pipeline and loads all models at the same time. With [`ModularPipeline`], creating and loading are two separate steps: [`~ModularPipeline.from_pretrained`] reads the configuration and knows where to load each component from, but doesn't actually load the model weights. You load the models later with [`~ModularPipeline.load_components`], which is where you pass loading arguments like `torch_dtype` and `quantization_config`.
+
+- **Two ways to create a pipeline.** You can use [`~ModularPipeline.from_pretrained`] with an existing diffusers model repository — it automatically maps to the default pipeline blocks and then converts to a [`ModularPipeline`] with no extra setup. Currently supported models include SDXL, Wan, Qwen, Z-Image, Flux, and Flux2. You can also assemble your own pipeline from [`ModularPipelineBlocks`] and convert it with the [`~ModularPipelineBlocks.init_pipeline`] method (see [Creating a pipeline](#creating-a-pipeline) for more details).
+
+- **Running the pipeline is the same.** Once loaded, you call the pipeline with the same arguments you're used to. A single [`ModularPipeline`] can support multiple workflows (text-to-image, image-to-image, inpainting, etc.) when the pipeline blocks use [`AutoPipelineBlocks`](./auto_pipeline_blocks) to automatically select the workflow based on your inputs.
+
+Below are complete examples for text-to-image, image-to-image, and inpainting with SDXL.
 
 <hfoptions id="example">
 <hfoption id="text-to-image">
 
 ```py
 import torch
-from diffusers.modular_pipelines import SequentialPipelineBlocks
-from diffusers.modular_pipelines.stable_diffusion_xl import TEXT2IMAGE_BLOCKS
-
-blocks = SequentialPipelineBlocks.from_blocks_dict(TEXT2IMAGE_BLOCKS)
-
-modular_repo_id = "YiYiXu/modular-loader-t2i-0704"
-pipeline = blocks.init_pipeline(modular_repo_id)
+from diffusers import ModularPipeline
 
+pipeline = ModularPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0")
 pipeline.load_components(torch_dtype=torch.float16)
 pipeline.to("cuda")
 
-image = pipeline(prompt="Astronaut in a jungle, cold color palette, muted colors, detailed, 8k", output="images")[0]
+image = pipeline(prompt="Astronaut in a jungle, cold color palette, muted colors, detailed, 8k").images[0]
 image.save("modular_t2i_out.png")
 ```
 
@@ -41,21 +42,17 @@ image.save("modular_t2i_out.png")
 
 ```py
 import torch
-from diffusers.modular_pipelines import SequentialPipelineBlocks
-from diffusers.modular_pipelines.stable_diffusion_xl import IMAGE2IMAGE_BLOCKS
-
-blocks = SequentialPipelineBlocks.from_blocks_dict(IMAGE2IMAGE_BLOCKS)
-
-modular_repo_id = "YiYiXu/modular-loader-t2i-0704"
-pipeline = blocks.init_pipeline(modular_repo_id)
+from diffusers import ModularPipeline
+from diffusers.utils import load_image
 
+pipeline = ModularPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0")
 pipeline.load_components(torch_dtype=torch.float16)
 pipeline.to("cuda")
 
 url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/sdxl-text2img.png"
 init_image = load_image(url)
 prompt = "a dog catching a frisbee in the jungle"
-image = pipeline(prompt=prompt, image=init_image, strength=0.8, output="images")[0]
+image = pipeline(prompt=prompt, image=init_image, strength=0.8).images[0]
 image.save("modular_i2i_out.png")
 ```
 
@@ -64,15 +61,10 @@ image.save("modular_i2i_out.png")
 
 ```py
 import torch
-from diffusers.modular_pipelines import SequentialPipelineBlocks
-from diffusers.modular_pipelines.stable_diffusion_xl import INPAINT_BLOCKS
+from diffusers import ModularPipeline
 from diffusers.utils import load_image
 
-blocks = SequentialPipelineBlocks.from_blocks_dict(INPAINT_BLOCKS)
-
-modular_repo_id = "YiYiXu/modular-loader-t2i-0704"
-pipeline = blocks.init_pipeline(modular_repo_id)
-
+pipeline = ModularPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0")
 pipeline.load_components(torch_dtype=torch.float16)
 pipeline.to("cuda")
 
@@ -83,276 +75,353 @@ init_image = load_image(img_url)
 mask_image = load_image(mask_url)
 
 prompt = "A deep sea diver floating"
-image = pipeline(prompt=prompt, image=init_image, mask_image=mask_image, strength=0.85, output="images")[0]
-image.save("moduar_inpaint_out.png")
+image = pipeline(prompt=prompt, image=init_image, mask_image=mask_image, strength=0.85).images[0]
+image.save("modular_inpaint_out.png")
 ```
 
 </hfoption>
 </hfoptions>
 
-This guide will show you how to create a [`ModularPipeline`] and manage the components in it.
-
-## Adding blocks
-
-Blocks are [`InsertableDict`] objects that can be inserted at specific positions, providing a flexible way to mix-and-match blocks.
-
-Use [`~modular_pipelines.modular_pipeline_utils.InsertableDict.insert`] on either the block class or `sub_blocks` attribute to add a block.
-
-```py
-# BLOCKS is dict of block classes, you need to add class to it
-BLOCKS.insert("block_name", BlockClass, index)
-# sub_blocks attribute contains instance, add a block instance to the  attribute
-t2i_blocks.sub_blocks.insert("block_name", block_instance, index)
-```
-
-Use [`~modular_pipelines.modular_pipeline_utils.InsertableDict.pop`] on either the block class or `sub_blocks` attribute to remove a block.
-
-```py
-# remove a block class from preset
-BLOCKS.pop("text_encoder")
-# split out a block instance on its own
-text_encoder_block = t2i_blocks.sub_blocks.pop("text_encoder")
-```
-
-Swap blocks by setting the existing block to the new block.
-
-```py
-# Replace block class in preset
-BLOCKS["prepare_latents"] = CustomPrepareLatents
-# Replace in sub_blocks attribute using an block instance
-t2i_blocks.sub_blocks["prepare_latents"] = CustomPrepareLatents()
-```
+This guide will show you how to create a [`ModularPipeline`], manage the components in it, and run it.
 
 ## Creating a pipeline
 
-There are two ways to create a [`ModularPipeline`]. Assemble and create a pipeline from [`ModularPipelineBlocks`] or load an existing pipeline with [`~ModularPipeline.from_pretrained`].
+There are two ways to create a [`ModularPipeline`]. Assemble and create a pipeline from [`ModularPipelineBlocks`] with [`~ModularPipelineBlocks.init_pipeline`], or load an existing pipeline with [`~ModularPipeline.from_pretrained`].
 
-You should also initialize a [`ComponentsManager`] to handle device placement and memory and component management.
+You can also initialize a [`ComponentsManager`](./components_manager) to handle device placement and memory management. If you don't need automatic offloading, you can skip this and move the pipeline to your device manually with `pipeline.to("cuda")`.
 
 > [!TIP]
 > Refer to the [ComponentsManager](./components_manager) doc for more details about how it can help manage components across different workflows.
 
-<hfoptions id="create">
-<hfoption id="ModularPipelineBlocks">
+### init_pipeline
 
-Use the [`~ModularPipelineBlocks.init_pipeline`] method to create a [`ModularPipeline`] from the component and configuration specifications. This method loads the *specifications* from a `modular_model_index.json` file, but it doesn't load the *models* yet.
+[`~ModularPipelineBlocks.init_pipeline`] converts any [`ModularPipelineBlocks`] into a [`ModularPipeline`].
+
+Let's define a minimal block to see how it works:
 
 ```py
-from diffusers import ComponentsManager
-from diffusers.modular_pipelines import SequentialPipelineBlocks
-from diffusers.modular_pipelines.stable_diffusion_xl import TEXT2IMAGE_BLOCKS
+from transformers import CLIPTextModel
+from diffusers.modular_pipelines import (
+    ComponentSpec,
+    ModularPipelineBlocks,
+    PipelineState,
+)
 
-t2i_blocks = SequentialPipelineBlocks.from_blocks_dict(TEXT2IMAGE_BLOCKS)
+class MyBlock(ModularPipelineBlocks):
+    @property
+    def expected_components(self):
+        return [
+            ComponentSpec(
+                name="text_encoder",
+                type_hint=CLIPTextModel,
+                pretrained_model_name_or_path="openai/clip-vit-large-patch14",
+            ),
+        ]
 
-modular_repo_id = "YiYiXu/modular-loader-t2i-0704"
-components = ComponentsManager()
-t2i_pipeline = t2i_blocks.init_pipeline(modular_repo_id, components_manager=components)
+    def __call__(self, components, state: PipelineState) -> PipelineState:
+        return components, state
 ```
 
-</hfoption>
-<hfoption id="from_pretrained">
+Call [`~ModularPipelineBlocks.init_pipeline`] to convert it into a pipeline. The `blocks` attribute on the pipeline is the blocks it was created from — it determines the expected inputs, outputs, and computation logic.
 
-The [`~ModularPipeline.from_pretrained`] method creates a [`ModularPipeline`] from a modular repository on the Hub.
+```py
+block = MyBlock()
+pipe = block.init_pipeline()
+pipe.blocks
+```
+
+```
+MyBlock {
+  "_class_name": "MyBlock",
+  "_diffusers_version": "0.37.0.dev0"
+}
+```
+
+> [!WARNING]
+> Blocks are mutable — you can freely add, remove, or swap blocks before creating a pipeline. However, once a pipeline is created, modifying `pipeline.blocks` won't affect the pipeline because it returns a copy. If you want a different block structure, create a new pipeline after modifying the blocks.
+
+When you call [`~ModularPipelineBlocks.init_pipeline`] without a repository, it uses the `pretrained_model_name_or_path` defined in the block's [`ComponentSpec`] to determine where to load each component from. Printing the pipeline shows the component loading configuration.
+
+```py
+pipe
+ModularPipeline {
+  "_blocks_class_name": "MyBlock",
+  "_class_name": "ModularPipeline",
+  "_diffusers_version": "0.37.0.dev0",
+  "text_encoder": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "openai/clip-vit-large-patch14",
+      "revision": null,
+      "subfolder": "",
+      "type_hint": [
+        "transformers",
+        "CLIPTextModel"
+      ],
+      "variant": null
+    }
+  ]
+}
+```
+
+If you pass a repository to [`~ModularPipelineBlocks.init_pipeline`], it overrides the loading path by matching your block's components against the pipeline config in that repository (`model_index.json` or `modular_model_index.json`).
+
+In the example below, the `pretrained_model_name_or_path` will be updated to `"stabilityai/stable-diffusion-xl-base-1.0"`.
+
+```py
+pipe = block.init_pipeline("stabilityai/stable-diffusion-xl-base-1.0")
+pipe
+ModularPipeline {
+  "_blocks_class_name": "MyBlock",
+  "_class_name": "ModularPipeline",
+  "_diffusers_version": "0.37.0.dev0",
+  "text_encoder": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "stabilityai/stable-diffusion-xl-base-1.0",
+      "revision": null,
+      "subfolder": "text_encoder",
+      "type_hint": [
+        "transformers",
+        "CLIPTextModel"
+      ],
+      "variant": null
+    }
+  ]
+}
+```
+
+If a component in your block doesn't exist in the repository, it remains `null` and is skipped during [`~ModularPipeline.load_components`].
+
+### from_pretrained
+
+[`~ModularPipeline.from_pretrained`] is a convenient way to create a [`ModularPipeline`] without defining blocks yourself.
+
+It works with three types of repositories.
+
+**A regular diffusers repository.** Pass any supported model repository and it automatically maps to the default pipeline blocks. Currently supported models include SDXL, Wan, Qwen, Z-Image, Flux, and Flux2.
 
 ```py
 from diffusers import ModularPipeline, ComponentsManager
 
 components = ComponentsManager()
-pipeline = ModularPipeline.from_pretrained("YiYiXu/modular-loader-t2i-0704", components_manager=components)
+pipeline = ModularPipeline.from_pretrained(
+    "stabilityai/stable-diffusion-xl-base-1.0", components_manager=components
+)
 ```
 
-Add the `trust_remote_code` argument to load a custom [`ModularPipeline`].
+**A modular repository.** These repositories contain a `modular_model_index.json` that specifies where to load each component from — the components can come from different repositories and the modular repository itself may not contain any model weights. For example, [diffusers/flux2-bnb-4bit-modular](https://huggingface.co/diffusers/flux2-bnb-4bit-modular) loads a quantized transformer from one repository and the remaining components from another. See [Modular repository](#modular-repository) for more details on the format.
 
 ```py
 from diffusers import ModularPipeline, ComponentsManager
 
 components = ComponentsManager()
-modular_repo_id = "YiYiXu/modular-diffdiff-0704"
-diffdiff_pipeline = ModularPipeline.from_pretrained(modular_repo_id, trust_remote_code=True, components_manager=components)
+pipeline = ModularPipeline.from_pretrained(
+    "diffusers/flux2-bnb-4bit-modular", components_manager=components
+)
 ```
 
-</hfoption>
-</hfoptions>
+**A modular repository with custom code.** Some repositories include custom pipeline blocks alongside the loading configuration. Add `trust_remote_code=True` to load them. See [Custom blocks](./custom_blocks) for how to create your own.
+
+```py
+from diffusers import ModularPipeline, ComponentsManager
+
+components = ComponentsManager()
+pipeline = ModularPipeline.from_pretrained(
+    "diffusers/Florence2-image-Annotator", trust_remote_code=True, components_manager=components
+)
+```
 
 ## Loading components
 
-A [`ModularPipeline`] doesn't automatically instantiate with components. It only loads the configuration and component specifications. You can load all components with [`~ModularPipeline.load_components`] or only load specific components with [`~ModularPipeline.load_components`].
+A [`ModularPipeline`] doesn't automatically instantiate with components. It only loads the configuration and component specifications. You can load components with [`~ModularPipeline.load_components`].
 
-<hfoptions id="load">
-<hfoption id="load_components">
+This will load all the components that have a valid loading spec.
 
 ```py
 import torch
 
-t2i_pipeline.load_components(torch_dtype=torch.float16)
-t2i_pipeline.to("cuda")
+pipeline.load_components(torch_dtype=torch.float16)
 ```
 
-</hfoption>
-<hfoption id="load_components">
-
-The example below only loads the UNet and VAE.
+You can also load specific components by name. The example below only loads the text_encoder.
 
 ```py
-import torch
-
-t2i_pipeline.load_components(names=["unet", "vae"], torch_dtype=torch.float16)
+pipeline.load_components(names=["text_encoder"], torch_dtype=torch.float16)
 ```
 
-</hfoption>
-</hfoptions>
-
-Print the pipeline to inspect the loaded pretrained components.
+After loading, printing the pipeline shows which components are loaded — the first two fields change from `null` to the component's library and class.
 
 ```py
-t2i_pipeline
+pipeline
 ```
 
-This should match the `modular_model_index.json` file from the modular repository a pipeline is initialized from. If a pipeline doesn't need a component, it won't be included even if it exists in the modular repository.
-
-To modify where components are loaded from, edit the `modular_model_index.json` file in the repository and change it to your desired loading path. The example below loads a UNet from a different repository.
-
-```json
-# original
-"unet": [
-  null, null,
-  {
-    "repo": "stabilityai/stable-diffusion-xl-base-1.0",
-    "subfolder": "unet",
-    "variant": "fp16"
-  }
+```
+# text_encoder is loaded - shows library and class
+"text_encoder": [
+  "transformers",
+  "CLIPTextModel",
+  { ... }
 ]
 
-# modified
+# unet is not loaded yet - still null
 "unet": [
-  null, null,
-  {
-    "repo": "RunDiffusion/Juggernaut-XL-v9",
-    "subfolder": "unet",
-    "variant": "fp16"
-  }
+  null,
+  null,
+  { ... }
 ]
 ```
 
-### Component loading status
-
-The pipeline properties below provide more information about which components are loaded.
-
-Use `component_names` to return all expected components.
+Loading keyword arguments like `torch_dtype`, `variant`, `revision`, and `quantization_config` are passed through to `from_pretrained()` for each component. You can pass a single value to apply to all components, or a dict to set per-component values.
 
 ```py
-t2i_pipeline.component_names
-['text_encoder', 'text_encoder_2', 'tokenizer', 'tokenizer_2', 'guider', 'scheduler', 'unet', 'vae', 'image_processor']
+# apply bfloat16 to all components
+pipeline.load_components(torch_dtype=torch.bfloat16)
+
+# different dtypes per component
+pipeline.load_components(torch_dtype={"transformer": torch.bfloat16, "default": torch.float32})
 ```
 
-Use `null_component_names` to return components that aren't loaded yet. Load these components with [`~ModularPipeline.from_pretrained`].
-
-```py
-t2i_pipeline.null_component_names
-['text_encoder', 'text_encoder_2', 'tokenizer', 'tokenizer_2', 'scheduler']
-```
-
-Use `pretrained_component_names` to return components that will be loaded from pretrained models.
-
-```py
-t2i_pipeline.pretrained_component_names
-['text_encoder', 'text_encoder_2', 'tokenizer', 'tokenizer_2', 'scheduler', 'unet', 'vae']
-```
-
-Use `config_component_names` to return components that are created with the default config (not loaded from a modular repository). Components from a config aren't included because they are already initialized during pipeline creation. This is why they aren't listed in `null_component_names`.
-
-```py
-t2i_pipeline.config_component_names
-['guider', 'image_processor']
-```
+Note that [`~ModularPipeline.load_components`] only loads components that haven't been loaded yet and have a valid loading spec. This means if you've already set a component on the pipeline, calling [`~ModularPipeline.load_components`] again won't reload it.
 
 ## Updating components
 
-Components may be updated depending on whether it is a *pretrained component* or a *config component*.
+[`~ModularPipeline.update_components`] replaces a component on the pipeline with a new one. When a component is updated, the loading specifications are also updated in the pipeline config and [`~ModularPipeline.load_components`] will skip it on subsequent calls.
 
-> [!WARNING]
-> A component may change from pretrained to config when updating a component. The component type is initially defined in a block's `expected_components` field.
+### From AutoModel
 
-A pretrained component is updated with [`ComponentSpec`] whereas a config component is updated by eihter passing the object directly or with [`ComponentSpec`].
-
-The [`ComponentSpec`] shows `default_creation_method="from_pretrained"` for a pretrained component shows `default_creation_method="from_config` for a config component.
-
-To update a pretrained component, create a [`ComponentSpec`] with the name of the component and where to load it from. Use the [`~ComponentSpec.load`] method to load the component.
+You can pass a model object loaded with `AutoModel.from_pretrained()`. Models loaded this way are automatically tagged with their loading information.
 
 ```py
-from diffusers import ComponentSpec, UNet2DConditionModel
+from diffusers import AutoModel
 
-unet_spec = ComponentSpec(name="unet",type_hint=UNet2DConditionModel, repo="stabilityai/stable-diffusion-xl-base-1.0", subfolder="unet", variant="fp16")
-unet = unet_spec.load(torch_dtype=torch.float16)
-```
-
-The [`~ModularPipeline.update_components`] method replaces the component with a new one.
-
-```py
-t2i_pipeline.update_components(unet=unet2)
-```
-
-When a component is updated, the loading specifications are also updated in the pipeline config.
-
-### Component extraction and modification
-
-When you use [`~ComponentSpec.load`], the new component maintains its loading specifications. This makes it possible to extract the specification and recreate the component.
-
-```py
-spec = ComponentSpec.from_component("unet", unet2)
-spec
-ComponentSpec(name='unet', type_hint=<class 'diffusers.models.unets.unet_2d_condition.UNet2DConditionModel'>, description=None, config=None, repo='stabilityai/stable-diffusion-xl-base-1.0', subfolder='unet', variant='fp16', revision=None, default_creation_method='from_pretrained')
-unet2_recreated = spec.load(torch_dtype=torch.float16)
-```
-
-The [`~ModularPipeline.get_component_spec`] method gets a copy of the current component specification to modify or update.
-
-```py
-unet_spec = t2i_pipeline.get_component_spec("unet")
-unet_spec
-ComponentSpec(
-    name='unet',
-    type_hint=<class 'diffusers.models.unets.unet_2d_condition.UNet2DConditionModel'>,
-    pretrained_model_name_or_path='RunDiffusion/Juggernaut-XL-v9',
-    subfolder='unet',
-    variant='fp16',
-    default_creation_method='from_pretrained'
+unet = AutoModel.from_pretrained(
+    "RunDiffusion/Juggernaut-XL-v9", subfolder="unet", variant="fp16", torch_dtype=torch.float16
 )
+pipeline.update_components(unet=unet)
+```
+
+### From ComponentSpec
+
+Use [`~ModularPipeline.get_component_spec`] to get a copy of the current component specification, modify it, and load a new component.
+
+```py
+unet_spec = pipeline.get_component_spec("unet")
 
 # modify to load from a different repository
-unet_spec.pretrained_model_name_or_path = "stabilityai/stable-diffusion-xl-base-1.0"
+unet_spec.pretrained_model_name_or_path = "RunDiffusion/Juggernaut-XL-v9"
 
-# load component with modified spec
+# load and update
 unet = unet_spec.load(torch_dtype=torch.float16)
+pipeline.update_components(unet=unet)
 ```
 
+You can also create a [`ComponentSpec`] from scratch.
+
+Not all components are loaded from pretrained weights — some are created from a config (listed under `pipeline.config_component_names`). For these, use [`~ComponentSpec.create`] instead of [`~ComponentSpec.load`].
+
+```py
+guider_spec = pipeline.get_component_spec("guider")
+guider_spec.config = {"guidance_scale": 5.0}
+guider = guider_spec.create()
+pipeline.update_components(guider=guider)
+```
+
+Or simply pass the object directly.
+
+```py
+from diffusers.guiders import ClassifierFreeGuidance
+
+guider = ClassifierFreeGuidance(guidance_scale=5.0)
+pipeline.update_components(guider=guider)
+```
+
+See the [Guiders](./guiders) guide for more details on available guiders and how to configure them.
+
+## Splitting a pipeline into stages
+
+Since blocks are composable, you can take a pipeline apart and reconstruct it into separate pipelines for each stage. The example below shows how we can separate the text encoder block from the rest of the pipeline, so you can encode the prompt independently and pass the embeddings to the main pipeline.
+
+```py
+from diffusers import ModularPipeline, ComponentsManager
+import torch
+
+device = "cuda"
+dtype = torch.bfloat16
+repo_id = "black-forest-labs/FLUX.2-klein-4B"
+
+# get the blocks and separate out the text encoder
+blocks = ModularPipeline.from_pretrained(repo_id).blocks
+text_block = blocks.sub_blocks.pop("text_encoder")
+
+# use ComponentsManager to handle offloading across multiple pipelines
+manager = ComponentsManager()
+manager.enable_auto_cpu_offload(device=device)
+
+# create separate pipelines for each stage
+text_encoder_pipeline = text_block.init_pipeline(repo_id, components_manager=manager)
+pipeline = blocks.init_pipeline(repo_id, components_manager=manager)
+
+# encode text
+text_encoder_pipeline.load_components(torch_dtype=dtype)
+text_embeddings = text_encoder_pipeline(prompt="a cat").get_by_kwargs("denoiser_input_fields")
+
+# denoise and decode
+pipeline.load_components(torch_dtype=dtype)
+output = pipeline(
+    **text_embeddings,
+    num_inference_steps=4,
+).images[0]
+```
+
+[`ComponentsManager`] handles memory across multiple pipelines. Unlike the offloading strategies in [`DiffusionPipeline`] that follow a fixed order, [`ComponentsManager`] makes offloading decisions dynamically each time a model forward pass runs, based on the current memory situation. This means it works regardless of how many pipelines you create or what order you run them in. See the [ComponentsManager](./components_manager) guide for more details.
+
+If pipeline stages share components (e.g., the same VAE used for encoding and decoding), you can use [`~ModularPipeline.update_components`] to pass an already-loaded component to another pipeline instead of loading it again.
+
 ## Modular repository
 
 A repository is required if the pipeline blocks use *pretrained components*. The repository supplies loading specifications and metadata.
 
-[`ModularPipeline`] specifically requires *modular repositories* (see [example repository](https://huggingface.co/YiYiXu/modular-diffdiff)) which are more flexible than a typical repository. It contains a `modular_model_index.json` file containing the following 3 elements.
+[`ModularPipeline`] works with regular diffusers repositories out of the box. However, you can also create a *modular repository* for more flexibility. A modular repository contains a `modular_model_index.json` file containing the following 3 elements.
 
-- `library` and `class` shows which library the component was loaded from and it's class. If `null`, the component hasn't been loaded yet.
+- `library` and `class` shows which library the component was loaded from and its class. If `null`, the component hasn't been loaded yet.
 - `loading_specs_dict` contains the information required to load the component such as the repository and subfolder it is loaded from.
 
-Unlike standard repositories, a modular repository can fetch components from different repositories based on the `loading_specs_dict`. Components don't need to exist in the same repository.
+The key advantage of a modular repository is that components can be loaded from different repositories. For example, [diffusers/flux2-bnb-4bit-modular](https://huggingface.co/diffusers/flux2-bnb-4bit-modular) loads a quantized transformer from `diffusers/FLUX.2-dev-bnb-4bit` while loading the remaining components from `black-forest-labs/FLUX.2-dev`.
 
-A modular repository may contain custom code for loading a [`ModularPipeline`]. This allows you to use specialized blocks that aren't native to Diffusers.
+To convert a regular diffusers repository into a modular one, create the pipeline using the regular repository, and then push to the Hub. The saved repository will contain a `modular_model_index.json` with all the loading specifications.
+
+```py
+from diffusers import ModularPipeline
+
+# load from a regular repo
+pipeline = ModularPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0")
+
+# push as a modular repository
+pipeline.save_pretrained("local/path", repo_id="my-username/sdxl-modular", push_to_hub=True)
+```
+
+A modular repository can also include custom pipeline blocks as Python code. This allows you to share specialized blocks that aren't native to Diffusers. For example, [diffusers/Florence2-image-Annotator](https://huggingface.co/diffusers/Florence2-image-Annotator) contains custom blocks alongside the loading configuration:
 
 ```
-modular-diffdiff-0704/
+Florence2-image-Annotator/
 ├── block.py                    # Custom pipeline blocks implementation
 ├── config.json                 # Pipeline configuration and auto_map
+├── mellon_config.json          # UI configuration for Mellon
 └── modular_model_index.json    # Component loading specifications
 ```
 
-The [config.json](https://huggingface.co/YiYiXu/modular-diffdiff-0704/blob/main/config.json) file contains an `auto_map` key that points to where a custom block is defined in `block.py`.
+The `config.json` file contains an `auto_map` key that tells [`ModularPipeline`] where to find the custom blocks:
 
 ```json
 {
-  "_class_name": "DiffDiffBlocks",
+  "_class_name": "Florence2AnnotatorBlocks",
   "auto_map": {
-    "ModularPipelineBlocks": "block.DiffDiffBlocks"
+    "ModularPipelineBlocks": "block.Florence2AnnotatorBlocks"
   }
 }
 ```
+
+Load custom code repositories with `trust_remote_code=True` as shown in [from_pretrained](#from_pretrained). See [Custom blocks](./custom_blocks) for how to create and share your own.

docs/source/en/modular_diffusers/pipeline_block.md

@@ -25,56 +25,42 @@ This guide will show you how to create a [`~modular_pipelines.ModularPipelineBlo
 
 A [`~modular_pipelines.ModularPipelineBlocks`] requires `inputs`, and `intermediate_outputs`.
 
-- `inputs` are values provided by a user and retrieved from the [`~modular_pipelines.PipelineState`]. This is useful because some workflows resize an image, but the original image is still required. The [`~modular_pipelines.PipelineState`] maintains the original image.
+- `inputs` are values a block reads from the [`~modular_pipelines.PipelineState`] to perform its computation. These can be values provided by a user (like a prompt or image) or values produced by a previous block (like encoded image_latents). 
 
     Use `InputParam` to define `inputs`.
 
-    ```py
-    from diffusers.modular_pipelines import InputParam
+```py
+class ImageEncodeStep(ModularPipelineBlocks):
+    ...
 
-    user_inputs = [
-        InputParam(name="image", type_hint="PIL.Image", description="raw input image to process")
-    ]
-    ```
+    @property
+    def inputs(self):
+        return [
+            InputParam(name="image", type_hint="PIL.Image", required=True, description="raw input image to process"),
+        ]
+    ...
+```
 
-- `intermediate_outputs` are new values created by a block and added to the [`~modular_pipelines.PipelineState`]. The `intermediate_outputs` are available as `inputs` for subsequent blocks or available as the final output from running the pipeline.
+-- `intermediate_outputs` are new values created by a block and added to the [`~modular_pipelines.PipelineState`]. The `intermediate_outputs` are available as `inputs` for subsequent blocks or available as the final output from running the pipeline.
 
     Use `OutputParam` to define `intermediate_outputs`.
 
-    ```py
-    from diffusers.modular_pipelines import OutputParam
+```py
+class ImageEncodeStep(ModularPipelineBlocks):
+    ...
 
-        user_intermediate_outputs = [
-        OutputParam(name="image_latents", description="latents representing the image")
-    ]
-    ```
+    @property
+    def intermediate_outputs(self):
+        return [
+            OutputParam(name="image_latents", description="latents representing the image"),
+        ]
+
+    ...
+```
 
 The intermediate inputs and outputs share data to connect blocks. They are accessible at any point, allowing you to track the workflow's progress.
 
-## Computation logic
-
-The computation a block performs is defined in the `__call__` method and it follows a specific structure.
-
-1. Retrieve the [`~modular_pipelines.BlockState`] to get a local view of the `inputs`
-2. Implement the computation logic on the `inputs`.
-3. Update [`~modular_pipelines.PipelineState`] to push changes from the local [`~modular_pipelines.BlockState`] back to the global [`~modular_pipelines.PipelineState`].
-4. Return the components and state which becomes available to the next block.
-
-```py
-def __call__(self, components, state):
-    # Get a local view of the state variables this block needs
-    block_state = self.get_block_state(state)
-
-    # Your computation logic here
-    # block_state contains all your inputs
-    # Access them like: block_state.image, block_state.processed_image
-
-    # Update the pipeline state with your updated block_states
-    self.set_block_state(state, block_state)
-    return components, state
-```
-
-### Components and configs
+## Components and configs
 
 The components and pipeline-level configs a block needs are specified in [`ComponentSpec`] and [`~modular_pipelines.ConfigSpec`].
 
@@ -82,24 +68,108 @@ The components and pipeline-level configs a block needs are specified in [`Compo
 - [`~modular_pipelines.ConfigSpec`] contains pipeline-level settings that control behavior across all blocks.
 
 ```py
-from diffusers import ComponentSpec, ConfigSpec
+class ImageEncodeStep(ModularPipelineBlocks):
+    ...
 
-expected_components = [
-    ComponentSpec(name="unet", type_hint=UNet2DConditionModel),
-    ComponentSpec(name="scheduler", type_hint=EulerDiscreteScheduler)
-]
+    @property
+    def expected_components(self):
+        return [
+            ComponentSpec(name="vae", type_hint=AutoencoderKL),
+        ]
 
-expected_config = [
-    ConfigSpec("force_zeros_for_empty_prompt", True)
-]
+    @property
+    def expected_configs(self):
+        return [
+            ConfigSpec("force_zeros_for_empty_prompt", True),
+        ]
+
+    ...
 ```
 
 When the blocks are converted into a pipeline, the components become available to the block as the first argument in `__call__`.
 
+## Computation logic
+
+The computation a block performs is defined in the `__call__` method and it follows a specific structure.
+
+1. Retrieve the [`~modular_pipelines.BlockState`] to get a local view of the `inputs`.
+2. Implement the computation logic on the `inputs`.
+3. Update [`~modular_pipelines.PipelineState`] to push changes from the local [`~modular_pipelines.BlockState`] back to the global [`~modular_pipelines.PipelineState`].
+4. Return the components and state which becomes available to the next block.
+
 ```py
-def __call__(self, components, state):
-    # Access components using dot notation
-    unet = components.unet
-    vae = components.vae
-    scheduler = components.scheduler
+class ImageEncodeStep(ModularPipelineBlocks):
+
+    def __call__(self, components, state):
+        # Get a local view of the state variables this block needs
+        block_state = self.get_block_state(state)
+
+        # Your computation logic here
+        # block_state contains all your inputs
+        # Access them like: block_state.image, block_state.processed_image
+
+        # Update the pipeline state with your updated block_states
+        self.set_block_state(state, block_state)
+        return components, state
 ```
+
+## Putting it all together
+
+Here is the complete block with all the pieces connected.
+
+```py
+from diffusers import ComponentSpec, AutoencoderKL
+from diffusers.modular_pipelines import InputParam, ModularPipelineBlocks, OutputParam
+
+
+class ImageEncodeStep(ModularPipelineBlocks):
+
+    @property
+    def description(self):
+        return "Encode an image into latent space."
+
+    @property
+    def expected_components(self):
+        return [
+            ComponentSpec(name="vae", type_hint=AutoencoderKL),
+        ]
+
+    @property
+    def inputs(self):
+        return [
+            InputParam(name="image", type_hint="PIL.Image", required=True, description="raw input image to process"),
+        ]
+
+    @property
+    def intermediate_outputs(self):
+        return [
+            OutputParam(name="image_latents", type_hint="torch.Tensor", description="latents representing the image"),
+        ]
+
+    def __call__(self, components, state):
+        block_state = self.get_block_state(state)
+        block_state.image_latents = components.vae.encode(block_state.image)
+        self.set_block_state(state, block_state)
+        return components, state
+```
+
+Every block has a `doc` property that is automatically generated from the properties you defined above. It provides a summary of the block's description, components, inputs, and outputs.
+
+```py
+block = ImageEncoderStep()
+print(block.doc)
+class ImageEncodeStep
+
+  Encode an image into latent space.
+
+  Components:
+      vae (`AutoencoderKL`)
+
+  Inputs:
+      image (`PIL.Image`):
+          raw input image to process
+
+  Outputs:
+      image_latents (`torch.Tensor`):
+          latents representing the image
+```

docs/source/en/modular_diffusers/quickstart.md

@@ -39,17 +39,44 @@ image
 [`~ModularPipeline.from_pretrained`] uses lazy loading - it reads the configuration to learn where to load each component from, but doesn't actually load the model weights until you call [`~ModularPipeline.load_components`]. This gives you control over when and how components are loaded.
 
 > [!TIP]
-> [`ComponentsManager`] with `enable_auto_cpu_offload` automatically moves models between CPU and GPU as needed, reducing memory usage for large models like Qwen-Image. Learn more in the [ComponentsManager](./components_manager) guide.
+> `ComponentsManager` with `enable_auto_cpu_offload` automatically moves models between CPU and GPU as needed, reducing memory usage for large models like Qwen-Image. Learn more in the [ComponentsManager](./components_manager) guide.
+>
+> If you don't need offloading, simply remove the `components_manager` argument and move the pipeline to your device manually with `pipe.to("cuda")`.
 
 Learn more about creating and loading pipelines in the [Creating a pipeline](https://huggingface.co/docs/diffusers/modular_diffusers/modular_pipeline#creating-a-pipeline) and [Loading components](https://huggingface.co/docs/diffusers/modular_diffusers/modular_pipeline#loading-components) guides.
 
 ## Understand the structure
 
-A [`ModularPipeline`] has two parts:
-- **State**: the loaded components (models, schedulers, processors) and configuration
-- **Definition**: the [`ModularPipelineBlocks`] that specify inputs, outputs, expected components and computation logic
+A [`ModularPipeline`] has two parts: a **definition** (the blocks) and a **state** (the loaded components and configs).
 
-The blocks define *what* the pipeline does. Access them through `pipe.blocks`.
+Print the pipeline to see its state — the components and their loading status and configuration.
+```py
+print(pipe)
+```
+```
+QwenImageModularPipeline {
+  "_blocks_class_name": "QwenImageAutoBlocks",
+  "_class_name": "QwenImageModularPipeline",
+  "_diffusers_version": "0.37.0.dev0",
+  "transformer": [
+    "diffusers",
+    "QwenImageTransformer2DModel",
+    {
+      "pretrained_model_name_or_path": "Qwen/Qwen-Image",
+      "revision": null,
+      "subfolder": "transformer",
+      "type_hint": [
+        "diffusers",
+        "QwenImageTransformer2DModel"
+      ],
+      "variant": null
+    }
+  ],
+  ...
+}
+```
+
+Access the definition through `pipe.blocks` — this is the [`~modular_pipelines.ModularPipelineBlocks`] that defines the pipeline's workflows, inputs, outputs, and computation logic.
 ```py
 print(pipe.blocks)
 ```
@@ -87,7 +114,9 @@ The output returns:
 
 ### Workflows
 
-`QwenImageAutoBlocks` is a [`ConditionalPipelineBlocks`], so this pipeline supports multiple workflows and adapts its behavior based on the inputs you provide. For example, if you pass `image` to the pipeline, it runs an image-to-image workflow instead of text-to-image. Let's see this in action with an example.
+This pipeline supports multiple workflows and adapts its behavior based on the inputs you provide. For example, if you pass `image` to the pipeline, it runs an image-to-image workflow instead of text-to-image. Learn more about how this works under the hood in the [AutoPipelineBlocks](https://huggingface.co/docs/diffusers/modular_diffusers/auto_pipeline_blocks) guide.
+
+Let's see this in action with an example.
 ```py
 from diffusers.utils import load_image
 
@@ -99,20 +128,21 @@ image = pipe(
 ).images[0]
 ```
 
-Use `get_workflow()` to extract the blocks for a specific workflow. Pass the workflow name (e.g., `"image2image"`, `"inpainting"`, `"controlnet_text2image"`) to get only the blocks relevant to that workflow.
+Use `get_workflow()` to extract the blocks for a specific workflow. Pass the workflow name (e.g., `"image2image"`, `"inpainting"`, `"controlnet_text2image"`) to get only the blocks relevant to that workflow. This is useful when you want to customize or debug a specific workflow.
 ```py
 img2img_blocks = pipe.blocks.get_workflow("image2image")
 ```
 
-Conditional blocks are convenient for users, but their conditional logic adds complexity when customizing or debugging. Extracting a workflow gives you the specific blocks relevant to your workflow, making it easier to work with. Learn more in the [AutoPipelineBlocks](https://huggingface.co/docs/diffusers/modular_diffusers/auto_pipeline_blocks) guide.
 
 ### Sub-blocks
 
 Blocks can contain other blocks. `pipe.blocks` gives you the top-level block definition (here, `QwenImageAutoBlocks`), while `sub_blocks` lets you access the smaller blocks inside it.
 
-`QwenImageAutoBlocks` is composed of: `text_encoder`, `vae_encoder`, `controlnet_vae_encoder`, `denoise`, and `decode`. Access them through the `sub_blocks` property.
+`QwenImageAutoBlocks` is composed of: `text_encoder`, `vae_encoder`, `controlnet_vae_encoder`, `denoise`, and `decode`.
 
-The `doc` property is useful for seeing the full documentation of any block, including its inputs, outputs, and components.
+These sub-blocks run one after another and data flows linearly from one block to the next — each block's `intermediate_outputs` become available as `inputs` to the next block. This is how [`SequentialPipelineBlocks`](./sequential_pipeline_blocks) work.
+
+You can access them through the `sub_blocks` property. The `doc` property is useful for seeing the full documentation of any block, including its inputs, outputs, and components.
 ```py
 vae_encoder_block = pipe.blocks.sub_blocks["vae_encoder"]
 print(vae_encoder_block.doc)
@@ -165,7 +195,7 @@ class CannyBlock
           Canny map for input image
 ```
 
-UUse `get_workflow` to extract the ControlNet workflow from [`QwenImageAutoBlocks`].
+Use `get_workflow` to extract the ControlNet workflow from [`QwenImageAutoBlocks`].
 ```py
 # Get the controlnet workflow that we want to work with
 blocks = pipe.blocks.get_workflow("controlnet_text2image")
@@ -182,9 +212,8 @@ class SequentialPipelineBlocks
       ...
 ```
 
-The extracted workflow is a [`SequentialPipelineBlocks`](./sequential_pipeline_blocks) - a multi-block type where blocks run one after another and data flows linearly from one block to the next. Each block's `intermediate_outputs` become available as `inputs` to subsequent blocks.
 
-Currently this workflow requires `control_image` as input. Let's insert the canny block at the beginning so the pipeline accepts a regular image instead.
+The extracted workflow is a [`SequentialPipelineBlocks`](./sequential_pipeline_blocks) and it currently requires `control_image` as input. Let's insert the canny block at the beginning so the pipeline accepts a regular image instead.
 ```py
 # Insert canny at the beginning
 blocks.sub_blocks.insert("canny", canny_block, 0)
@@ -211,7 +240,7 @@ class SequentialPipelineBlocks
 
 Now the pipeline takes `image` as input instead of `control_image`. Because blocks in a sequence share data automatically, the canny block's output (`control_image`) flows to the denoise block that needs it, and the canny block's input (`image`) becomes a pipeline input since no earlier block provides it.
 
-Create a pipeline from the modified blocks and load a ControlNet model.
+Create a pipeline from the modified blocks and load a ControlNet model. The ControlNet isn't part of the original model repository, so we load it separately and add it with [`~ModularPipeline.update_components`].
 ```py
 pipeline = blocks.init_pipeline("Qwen/Qwen-Image", components_manager=manager)
 
@@ -241,6 +270,16 @@ output
 ## Next steps
 
 <hfoptions id="next">
+<hfoption id="Learn the basics">
+
+Understand the core building blocks of Modular Diffusers:
+
+- [ModularPipelineBlocks](./pipeline_block): The basic unit for defining a step in a pipeline.
+- [SequentialPipelineBlocks](./sequential_pipeline_blocks): Chain blocks to run in sequence.
+- [AutoPipelineBlocks](./auto_pipeline_blocks): Create pipelines that support multiple workflows.
+- [States](./modular_diffusers_states): How data is shared between blocks.
+
+</hfoption>
 <hfoption id="Build custom blocks">
 
 Learn how to create your own blocks with custom logic in the [Building Custom Blocks](./custom_blocks) guide.

docs/source/en/modular_diffusers/sequential_pipeline_blocks.md

@@ -91,23 +91,42 @@ class ImageEncoderBlock(ModularPipelineBlocks):
 </hfoption>
 </hfoptions>
 
-Connect the two blocks by defining an [`InsertableDict`] to map the block names to the block instances. Blocks are executed in the order they're registered in `blocks_dict`.
-
-Use [`~modular_pipelines.SequentialPipelineBlocks.from_blocks_dict`] to create a [`~modular_pipelines.SequentialPipelineBlocks`].
+Connect the two blocks by defining a [~modular_pipelines.SequentialPipelineBlocks]. List the block instances in `block_classes` and their corresponding names in `block_names`. The blocks are executed in the order they appear in `block_classes`, and data flows from one block to the next through [~modular_pipelines.PipelineState].
 
 ```py
-from diffusers.modular_pipelines import SequentialPipelineBlocks, InsertableDict
+class ImageProcessingStep(SequentialPipelineBlocks):
+    """
+    # auto_docstring
+    """
+    model_name = "my_model"
+    block_classes = [InputBlock(), ImageEncoderBlock()]
+    block_names = ["input", "image_encoder"]
 
-blocks_dict = InsertableDict()
-blocks_dict["input"] = input_block
-blocks_dict["image_encoder"] = image_encoder_block
-
-blocks = SequentialPipelineBlocks.from_blocks_dict(blocks_dict)
+    @property
+    def description(self):
+        return (
+            "Process text prompts and images for the pipeline. It:\n"
+            " - Determines the batch size from the prompts.\n"
+            " - Encodes the image into latent space."
+        )
 ```
 
-Inspect the sub-blocks in [`~modular_pipelines.SequentialPipelineBlocks`] by calling `blocks`, and for more details about the inputs and outputs, access the `docs` attribute.
+When you create a [~modular_pipelines.SequentialPipelineBlocks], properties like `inputs`, `intermediate_outputs`, and `expected_components` are automatically aggregated from the sub-blocks, so there is no need to define them again.
+
+There are a few properties you should set:
+
+- `description`: We recommend adding a description for the assembled block to explain what the combined step does.
+- `model_name`: This is automatically derived from the sub-blocks but isn't always correct, so you may need to override it.
+- `outputs`: By default this is the same as `intermediate_outputs`, but you can manually set it to control which values appear in the doc. This is useful for showing only the final outputs instead of all intermediate values.
+
+These properties, together with the aggregated `inputs`, `intermediate_outputs`, and `expected_components`, are used to automatically generate the `doc` property.
+
+
+Inspect the sub-blocks through the `sub_blocks` property, and use `doc` for a full summary of the block's inputs, outputs, and components.
+
 
 ```py
+blocks = ImageProcessingStep()
 print(blocks)
 print(blocks.doc)
-```
+```

src/diffusers/modular_pipelines/modular_pipeline.py

@@ -1598,11 +1598,11 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
             else:
                 logger.warning(f"`blocks` is `None`, no default blocks class found for {self.__class__.__name__}")
 
-        self.blocks = blocks
+        self._blocks = blocks
         self._components_manager = components_manager
         self._collection = collection
-        self._component_specs = {spec.name: deepcopy(spec) for spec in self.blocks.expected_components}
-        self._config_specs = {spec.name: deepcopy(spec) for spec in self.blocks.expected_configs}
+        self._component_specs = {spec.name: deepcopy(spec) for spec in self._blocks.expected_components}
+        self._config_specs = {spec.name: deepcopy(spec) for spec in self._blocks.expected_configs}
 
         # update component_specs and config_specs based on modular_model_index.json
         if modular_config_dict is not None:
@@ -1649,7 +1649,9 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         for name, config_spec in self._config_specs.items():
             default_configs[name] = config_spec.default
         self.register_to_config(**default_configs)
-        self.register_to_config(_blocks_class_name=self.blocks.__class__.__name__ if self.blocks is not None else None)
+        self.register_to_config(
+            _blocks_class_name=self._blocks.__class__.__name__ if self._blocks is not None else None
+        )
 
     @property
     def default_call_parameters(self) -> Dict[str, Any]:
@@ -1658,7 +1660,7 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
             - Dictionary mapping input names to their default values
         """
         params = {}
-        for input_param in self.blocks.inputs:
+        for input_param in self._blocks.inputs:
             params[input_param.name] = input_param.default
         return params
 
@@ -1829,7 +1831,15 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         Returns:
             - The docstring of the pipeline blocks
         """
-        return self.blocks.doc
+        return self._blocks.doc
+
+    @property
+    def blocks(self) -> ModularPipelineBlocks:
+        """
+        Returns:
+            - A copy of the pipeline blocks
+        """
+        return deepcopy(self._blocks)
 
     def register_components(self, **kwargs):
         """
@@ -2057,58 +2067,29 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         - the `config` dict, which will be saved as `modular_model_index.json` during `save_pretrained`
 
         Args:
-            **kwargs: Component objects, ComponentSpec objects, or configuration values to update:
-                - Component objects: Only supports components we can extract specs using
-                  `ComponentSpec.from_component()` method i.e. components created with ComponentSpec.load() or
-                  ConfigMixin subclasses that aren't nn.Modules (e.g., `unet=new_unet, text_encoder=new_encoder`)
-                - ComponentSpec objects: Only supports default_creation_method == "from_config", will call create()
-                  method to create a new component (e.g., `guider=ComponentSpec(name="guider",
-                  type_hint=ClassifierFreeGuidance, config={...}, default_creation_method="from_config")`)
-                - Configuration values: Simple values to update configuration settings (e.g.,
-                  `requires_safety_checker=False`)
-
-        Raises:
-            ValueError: If a component object is not supported in ComponentSpec.from_component() method:
-                - nn.Module components without a valid `_diffusers_load_id` attribute
-                - Non-ConfigMixin components without a valid `_diffusers_load_id` attribute
+            **kwargs: Component objects or configuration values to update:
+                - Component objects: Models loaded with `AutoModel.from_pretrained()` or `ComponentSpec.load()`
+                are automatically tagged with loading information. ConfigMixin objects without weights (e.g.,
+                schedulers, guiders) can be passed directly.
+                - Configuration values: Simple values to update configuration settings
+                (e.g., `requires_safety_checker=False`)
 
         Examples:
             ```python
-            # Update multiple components at once
+            # Update pretrrained model
             pipeline.update_components(unet=new_unet_model, text_encoder=new_text_encoder)
 
             # Update configuration values
             pipeline.update_components(requires_safety_checker=False)
-
-            # Update both components and configs together
-            pipeline.update_components(unet=new_unet_model, requires_safety_checker=False)
-
-            # Update with ComponentSpec objects (from_config only)
-            pipeline.update_components(
-                guider=ComponentSpec(
-                    name="guider",
-                    type_hint=ClassifierFreeGuidance,
-                    config={"guidance_scale": 5.0},
-                    default_creation_method="from_config",
-                )
-            )
             ```
 
         Notes:
-            - Components with trained weights must be created using ComponentSpec.load(). If the component has not been
-              shared in huggingface hub and you don't have loading specs, you can upload it using `push_to_hub()`
-            - ConfigMixin objects without weights (e.g., schedulers, guiders) can be passed directly
-            - ComponentSpec objects with default_creation_method="from_pretrained" are not supported in
-              update_components()
+            - Components with trained weights should be loaded with `AutoModel.from_pretrained()` or
+            `ComponentSpec.load()` so that loading specs are preserved for serialization.
+            - ConfigMixin objects without weights (e.g., schedulers, guiders) can be passed directly.
         """
 
-        # extract component_specs_updates & config_specs_updates from `specs`
-        passed_component_specs = {
-            k: kwargs.pop(k) for k in self._component_specs if k in kwargs and isinstance(kwargs[k], ComponentSpec)
-        }
-        passed_components = {
-            k: kwargs.pop(k) for k in self._component_specs if k in kwargs and not isinstance(kwargs[k], ComponentSpec)
-        }
+        passed_components = {k: kwargs.pop(k) for k in self._component_specs if k in kwargs}
         passed_config_values = {k: kwargs.pop(k) for k in self._config_specs if k in kwargs}
 
         for name, component in passed_components.items():
@@ -2147,33 +2128,14 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         if len(kwargs) > 0:
             logger.warning(f"Unexpected keyword arguments, will be ignored: {kwargs.keys()}")
 
-        created_components = {}
-        for name, component_spec in passed_component_specs.items():
-            if component_spec.default_creation_method == "from_pretrained":
-                raise ValueError(
-                    "ComponentSpec object with default_creation_method == 'from_pretrained' is not supported in update_components() method"
-                )
-            created_components[name] = component_spec.create()
-            current_component_spec = self._component_specs[name]
-            # warn if type changed
-            if current_component_spec.type_hint is not None and not isinstance(
-                created_components[name], current_component_spec.type_hint
-            ):
-                logger.info(
-                    f"ModularPipeline.update_components: adding {name} with new type: {created_components[name].__class__.__name__}, previous type: {current_component_spec.type_hint.__name__}"
-                )
-            # update _component_specs based on the user passed component_spec
-            self._component_specs[name] = component_spec
-        self.register_components(**passed_components, **created_components)
+        self.register_components(**passed_components)
 
         config_to_register = {}
         for name, new_value in passed_config_values.items():
-            # e.g. requires_aesthetics_score = False
             self._config_specs[name].default = new_value
             config_to_register[name] = new_value
         self.register_to_config(**config_to_register)
 
-    # YiYi TODO: support map for additional from_pretrained kwargs
     def load_components(self, names: Optional[Union[List[str], str]] = None, **kwargs):
         """
         Load selected components from specs.
@@ -2565,7 +2527,7 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         )
 
     def set_progress_bar_config(self, **kwargs):
-        for sub_block_name, sub_block in self.blocks.sub_blocks.items():
+        for sub_block_name, sub_block in self._blocks.sub_blocks.items():
             if hasattr(sub_block, "set_progress_bar_config"):
                 sub_block.set_progress_bar_config(**kwargs)
 
@@ -2619,7 +2581,7 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
 
         # Add inputs to state, using defaults if not provided in the kwargs or the state
         # if same input already in the state, will override it if provided in the kwargs
-        for expected_input_param in self.blocks.inputs:
+        for expected_input_param in self._blocks.inputs:
             name = expected_input_param.name
             default = expected_input_param.default
             kwargs_type = expected_input_param.kwargs_type
@@ -2638,9 +2600,9 @@ class ModularPipeline(ConfigMixin, PushToHubMixin):
         # Run the pipeline
         with torch.no_grad():
             try:
-                _, state = self.blocks(self, state)
+                _, state = self._blocks(self, state)
             except Exception:
-                error_msg = f"Error in block: ({self.blocks.__class__.__name__}):\n"
+                error_msg = f"Error in block: ({self._blocks.__class__.__name__}):\n"
                 logger.error(error_msg)
                 raise
 

tests/models/transformers/test_models_transformer_flux.py

@@ -219,10 +219,6 @@ class TestFluxTransformerMemory(FluxTransformerTesterConfig, MemoryTesterMixin):
 class TestFluxTransformerTraining(FluxTransformerTesterConfig, TrainingTesterMixin):
     """Training tests for Flux Transformer."""
 
-    def test_gradient_checkpointing_is_applied(self):
-        expected_set = {"FluxTransformer2DModel"}
-        super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
-
 
 class TestFluxTransformerAttention(FluxTransformerTesterConfig, AttentionTesterMixin):
     """Attention processor tests for Flux Transformer."""

tests/models/transformers/test_models_transformer_flux2.py

@@ -13,88 +13,48 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import unittest
+
 import torch
 
-from diffusers import Flux2Transformer2DModel
-from diffusers.utils.torch_utils import randn_tensor
+from diffusers import Flux2Transformer2DModel, attention_backend
 
 from ...testing_utils import enable_full_determinism, torch_device
-from ..testing_utils import (
-    AttentionTesterMixin,
-    BaseModelTesterConfig,
-    BitsAndBytesTesterMixin,
-    ContextParallelTesterMixin,
-    GGUFCompileTesterMixin,
-    GGUFTesterMixin,
-    LoraHotSwappingForModelTesterMixin,
-    LoraTesterMixin,
-    MemoryTesterMixin,
-    ModelTesterMixin,
-    TorchAoTesterMixin,
-    TorchCompileTesterMixin,
-    TrainingTesterMixin,
-)
+from ..test_modeling_common import LoraHotSwappingForModelTesterMixin, ModelTesterMixin, TorchCompileTesterMixin
 
 
 enable_full_determinism()
 
 
-class Flux2TransformerTesterConfig(BaseModelTesterConfig):
-    @property
-    def model_class(self):
-        return Flux2Transformer2DModel
+class Flux2TransformerTests(ModelTesterMixin, unittest.TestCase):
+    model_class = Flux2Transformer2DModel
+    main_input_name = "hidden_states"
+    # We override the items here because the transformer under consideration is small.
+    model_split_percents = [0.7, 0.6, 0.6]
+
+    # Skip setting testing with default: AttnProcessor
+    uses_custom_attn_processor = True
 
     @property
-    def output_shape(self) -> tuple[int, int]:
+    def dummy_input(self):
+        return self.prepare_dummy_input()
+
+    @property
+    def input_shape(self):
         return (16, 4)
 
     @property
-    def input_shape(self) -> tuple[int, int]:
+    def output_shape(self):
         return (16, 4)
 
-    @property
-    def model_split_percents(self) -> list:
-        # We override the items here because the transformer under consideration is small.
-        return [0.7, 0.6, 0.6]
-
-    @property
-    def main_input_name(self) -> str:
-        return "hidden_states"
-
-    @property
-    def uses_custom_attn_processor(self) -> bool:
-        # Skip setting testing with default: AttnProcessor
-        return True
-
-    @property
-    def generator(self):
-        return torch.Generator("cpu").manual_seed(0)
-
-    def get_init_dict(self) -> dict[str, int | list[int]]:
-        return {
-            "patch_size": 1,
-            "in_channels": 4,
-            "num_layers": 1,
-            "num_single_layers": 1,
-            "attention_head_dim": 16,
-            "num_attention_heads": 2,
-            "joint_attention_dim": 32,
-            "timestep_guidance_channels": 256,  # Hardcoded in original code
-            "axes_dims_rope": [4, 4, 4, 4],
-        }
-
-    def get_dummy_inputs(self, height: int = 4, width: int = 4) -> dict[str, torch.Tensor]:
+    def prepare_dummy_input(self, height=4, width=4):
         batch_size = 1
         num_latent_channels = 4
         sequence_length = 48
         embedding_dim = 32
 
-        hidden_states = randn_tensor(
-            (batch_size, height * width, num_latent_channels), generator=self.generator, device=torch_device
-        )
-        encoder_hidden_states = randn_tensor(
-            (batch_size, sequence_length, embedding_dim), generator=self.generator, device=torch_device
-        )
+        hidden_states = torch.randn((batch_size, height * width, num_latent_channels)).to(torch_device)
+        encoder_hidden_states = torch.randn((batch_size, sequence_length, embedding_dim)).to(torch_device)
 
         t_coords = torch.arange(1)
         h_coords = torch.arange(height)
@@ -122,244 +82,81 @@ class Flux2TransformerTesterConfig(BaseModelTesterConfig):
             "guidance": guidance,
         }
 
+    def prepare_init_args_and_inputs_for_common(self):
+        init_dict = {
+            "patch_size": 1,
+            "in_channels": 4,
+            "num_layers": 1,
+            "num_single_layers": 1,
+            "attention_head_dim": 16,
+            "num_attention_heads": 2,
+            "joint_attention_dim": 32,
+            "timestep_guidance_channels": 256,  # Hardcoded in original code
+            "axes_dims_rope": [4, 4, 4, 4],
+        }
 
-class TestFlux2Transformer(Flux2TransformerTesterConfig, ModelTesterMixin):
-    pass
+        inputs_dict = self.dummy_input
+        return init_dict, inputs_dict
 
+    # TODO (Daniel, Sayak): We can remove this test.
+    def test_flux2_consistency(self, seed=0):
+        torch.manual_seed(seed)
+        init_dict, inputs_dict = self.prepare_init_args_and_inputs_for_common()
 
-class TestFlux2TransformerMemory(Flux2TransformerTesterConfig, MemoryTesterMixin):
-    """Memory optimization tests for Flux2 Transformer."""
+        torch.manual_seed(seed)
+        model = self.model_class(**init_dict)
+        # state_dict = model.state_dict()
+        # for key, param in state_dict.items():
+        #     print(f"{key} | {param.shape}")
+        # torch.save(state_dict, "/raid/daniel_gu/test_flux2_params/diffusers.pt")
+        model.to(torch_device)
+        model.eval()
 
+        with attention_backend("native"):
+            with torch.no_grad():
+                output = model(**inputs_dict)
 
-class TestFlux2TransformerTraining(Flux2TransformerTesterConfig, TrainingTesterMixin):
-    """Training tests for Flux2 Transformer."""
+                if isinstance(output, dict):
+                    output = output.to_tuple()[0]
+
+        self.assertIsNotNone(output)
+
+        # input & output have to have the same shape
+        input_tensor = inputs_dict[self.main_input_name]
+        expected_shape = input_tensor.shape
+        self.assertEqual(output.shape, expected_shape, "Input and output shapes do not match")
+
+        # Check against expected slice
+        # fmt: off
+        expected_slice = torch.tensor([-0.3662, 0.4844, 0.6334, -0.3497, 0.2162, 0.0188, 0.0521, -0.2061, -0.2041, -0.0342, -0.7107, 0.4797, -0.3280, 0.7059, -0.0849, 0.4416])
+        # fmt: on
+
+        flat_output = output.cpu().flatten()
+        generated_slice = torch.cat([flat_output[:8], flat_output[-8:]])
+        self.assertTrue(torch.allclose(generated_slice, expected_slice, atol=1e-4))
 
     def test_gradient_checkpointing_is_applied(self):
         expected_set = {"Flux2Transformer2DModel"}
         super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
 
 
-class TestFlux2TransformerAttention(Flux2TransformerTesterConfig, AttentionTesterMixin):
-    """Attention processor tests for Flux2 Transformer."""
+class Flux2TransformerCompileTests(TorchCompileTesterMixin, unittest.TestCase):
+    model_class = Flux2Transformer2DModel
+    different_shapes_for_compilation = [(4, 4), (4, 8), (8, 8)]
+
+    def prepare_init_args_and_inputs_for_common(self):
+        return Flux2TransformerTests().prepare_init_args_and_inputs_for_common()
+
+    def prepare_dummy_input(self, height, width):
+        return Flux2TransformerTests().prepare_dummy_input(height=height, width=width)
 
 
-class TestFlux2TransformerContextParallel(Flux2TransformerTesterConfig, ContextParallelTesterMixin):
-    """Context Parallel inference tests for Flux2 Transformer."""
+class Flux2TransformerLoRAHotSwapTests(LoraHotSwappingForModelTesterMixin, unittest.TestCase):
+    model_class = Flux2Transformer2DModel
+    different_shapes_for_compilation = [(4, 4), (4, 8), (8, 8)]
 
+    def prepare_init_args_and_inputs_for_common(self):
+        return Flux2TransformerTests().prepare_init_args_and_inputs_for_common()
 
-class TestFlux2TransformerLoRA(Flux2TransformerTesterConfig, LoraTesterMixin):
-    """LoRA adapter tests for Flux2 Transformer."""
-
-
-class TestFlux2TransformerLoRAHotSwap(Flux2TransformerTesterConfig, LoraHotSwappingForModelTesterMixin):
-    """LoRA hot-swapping tests for Flux2 Transformer."""
-
-    @property
-    def different_shapes_for_compilation(self):
-        return [(4, 4), (4, 8), (8, 8)]
-
-    def get_dummy_inputs(self, height: int = 4, width: int = 4) -> dict[str, torch.Tensor]:
-        """Override to support dynamic height/width for LoRA hotswap tests."""
-        batch_size = 1
-        num_latent_channels = 4
-        sequence_length = 48
-        embedding_dim = 32
-
-        hidden_states = randn_tensor(
-            (batch_size, height * width, num_latent_channels), generator=self.generator, device=torch_device
-        )
-        encoder_hidden_states = randn_tensor(
-            (batch_size, sequence_length, embedding_dim), generator=self.generator, device=torch_device
-        )
-
-        t_coords = torch.arange(1)
-        h_coords = torch.arange(height)
-        w_coords = torch.arange(width)
-        l_coords = torch.arange(1)
-        image_ids = torch.cartesian_prod(t_coords, h_coords, w_coords, l_coords)
-        image_ids = image_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        text_t_coords = torch.arange(1)
-        text_h_coords = torch.arange(1)
-        text_w_coords = torch.arange(1)
-        text_l_coords = torch.arange(sequence_length)
-        text_ids = torch.cartesian_prod(text_t_coords, text_h_coords, text_w_coords, text_l_coords)
-        text_ids = text_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        timestep = torch.tensor([1.0]).to(torch_device).expand(batch_size)
-        guidance = torch.tensor([1.0]).to(torch_device).expand(batch_size)
-
-        return {
-            "hidden_states": hidden_states,
-            "encoder_hidden_states": encoder_hidden_states,
-            "img_ids": image_ids,
-            "txt_ids": text_ids,
-            "timestep": timestep,
-            "guidance": guidance,
-        }
-
-
-class TestFlux2TransformerCompile(Flux2TransformerTesterConfig, TorchCompileTesterMixin):
-    @property
-    def different_shapes_for_compilation(self):
-        return [(4, 4), (4, 8), (8, 8)]
-
-    def get_dummy_inputs(self, height: int = 4, width: int = 4) -> dict[str, torch.Tensor]:
-        """Override to support dynamic height/width for compilation tests."""
-        batch_size = 1
-        num_latent_channels = 4
-        sequence_length = 48
-        embedding_dim = 32
-
-        hidden_states = randn_tensor(
-            (batch_size, height * width, num_latent_channels), generator=self.generator, device=torch_device
-        )
-        encoder_hidden_states = randn_tensor(
-            (batch_size, sequence_length, embedding_dim), generator=self.generator, device=torch_device
-        )
-
-        t_coords = torch.arange(1)
-        h_coords = torch.arange(height)
-        w_coords = torch.arange(width)
-        l_coords = torch.arange(1)
-        image_ids = torch.cartesian_prod(t_coords, h_coords, w_coords, l_coords)
-        image_ids = image_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        text_t_coords = torch.arange(1)
-        text_h_coords = torch.arange(1)
-        text_w_coords = torch.arange(1)
-        text_l_coords = torch.arange(sequence_length)
-        text_ids = torch.cartesian_prod(text_t_coords, text_h_coords, text_w_coords, text_l_coords)
-        text_ids = text_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        timestep = torch.tensor([1.0]).to(torch_device).expand(batch_size)
-        guidance = torch.tensor([1.0]).to(torch_device).expand(batch_size)
-
-        return {
-            "hidden_states": hidden_states,
-            "encoder_hidden_states": encoder_hidden_states,
-            "img_ids": image_ids,
-            "txt_ids": text_ids,
-            "timestep": timestep,
-            "guidance": guidance,
-        }
-
-
-class TestFlux2TransformerBitsAndBytes(Flux2TransformerTesterConfig, BitsAndBytesTesterMixin):
-    """BitsAndBytes quantization tests for Flux2 Transformer."""
-
-
-class TestFlux2TransformerTorchAo(Flux2TransformerTesterConfig, TorchAoTesterMixin):
-    """TorchAO quantization tests for Flux2 Transformer."""
-
-
-class TestFlux2TransformerGGUF(Flux2TransformerTesterConfig, GGUFTesterMixin):
-    """GGUF quantization tests for Flux2 Transformer."""
-
-    @property
-    def gguf_filename(self):
-        return "https://huggingface.co/unsloth/FLUX.2-dev-GGUF/blob/main/flux2-dev-Q2_K.gguf"
-
-    @property
-    def torch_dtype(self):
-        return torch.bfloat16
-
-    def get_dummy_inputs(self):
-        """Override to provide inputs matching the real FLUX2 model dimensions.
-
-        Flux2 defaults: in_channels=128, joint_attention_dim=15360
-        """
-        batch_size = 1
-        height = 64
-        width = 64
-        sequence_length = 512
-
-        hidden_states = randn_tensor(
-            (batch_size, height * width, 128), generator=self.generator, device=torch_device, dtype=self.torch_dtype
-        )
-        encoder_hidden_states = randn_tensor(
-            (batch_size, sequence_length, 15360), generator=self.generator, device=torch_device, dtype=self.torch_dtype
-        )
-
-        # Flux2 uses 4D image/text IDs (t, h, w, l)
-        t_coords = torch.arange(1)
-        h_coords = torch.arange(height)
-        w_coords = torch.arange(width)
-        l_coords = torch.arange(1)
-        image_ids = torch.cartesian_prod(t_coords, h_coords, w_coords, l_coords)
-        image_ids = image_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        text_t_coords = torch.arange(1)
-        text_h_coords = torch.arange(1)
-        text_w_coords = torch.arange(1)
-        text_l_coords = torch.arange(sequence_length)
-        text_ids = torch.cartesian_prod(text_t_coords, text_h_coords, text_w_coords, text_l_coords)
-        text_ids = text_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        timestep = torch.tensor([1.0]).to(torch_device, self.torch_dtype)
-        guidance = torch.tensor([3.5]).to(torch_device, self.torch_dtype)
-
-        return {
-            "hidden_states": hidden_states,
-            "encoder_hidden_states": encoder_hidden_states,
-            "img_ids": image_ids,
-            "txt_ids": text_ids,
-            "timestep": timestep,
-            "guidance": guidance,
-        }
-
-
-class TestFlux2TransformerGGUFCompile(Flux2TransformerTesterConfig, GGUFCompileTesterMixin):
-    """GGUF + compile tests for Flux2 Transformer."""
-
-    @property
-    def gguf_filename(self):
-        return "https://huggingface.co/unsloth/FLUX.2-dev-GGUF/blob/main/flux2-dev-Q2_K.gguf"
-
-    @property
-    def torch_dtype(self):
-        return torch.bfloat16
-
-    def get_dummy_inputs(self):
-        """Override to provide inputs matching the real FLUX2 model dimensions.
-
-        Flux2 defaults: in_channels=128, joint_attention_dim=15360
-        """
-        batch_size = 1
-        height = 64
-        width = 64
-        sequence_length = 512
-
-        hidden_states = randn_tensor(
-            (batch_size, height * width, 128), generator=self.generator, device=torch_device, dtype=self.torch_dtype
-        )
-        encoder_hidden_states = randn_tensor(
-            (batch_size, sequence_length, 15360), generator=self.generator, device=torch_device, dtype=self.torch_dtype
-        )
-
-        # Flux2 uses 4D image/text IDs (t, h, w, l)
-        t_coords = torch.arange(1)
-        h_coords = torch.arange(height)
-        w_coords = torch.arange(width)
-        l_coords = torch.arange(1)
-        image_ids = torch.cartesian_prod(t_coords, h_coords, w_coords, l_coords)
-        image_ids = image_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        text_t_coords = torch.arange(1)
-        text_h_coords = torch.arange(1)
-        text_w_coords = torch.arange(1)
-        text_l_coords = torch.arange(sequence_length)
-        text_ids = torch.cartesian_prod(text_t_coords, text_h_coords, text_w_coords, text_l_coords)
-        text_ids = text_ids.unsqueeze(0).expand(batch_size, -1, -1).to(torch_device)
-
-        timestep = torch.tensor([1.0]).to(torch_device, self.torch_dtype)
-        guidance = torch.tensor([3.5]).to(torch_device, self.torch_dtype)
-
-        return {
-            "hidden_states": hidden_states,
-            "encoder_hidden_states": encoder_hidden_states,
-            "img_ids": image_ids,
-            "txt_ids": text_ids,
-            "timestep": timestep,
-            "guidance": guidance,
-        }
+    def prepare_dummy_input(self, height, width):
+        return Flux2TransformerTests().prepare_dummy_input(height=height, width=width)