Expose max_area for condition images in Flux2 pipelines

The condition/reference image downscale threshold was hardcoded to
1024*1024 in Flux2Pipeline, Flux2KleinPipeline and Flux2KleinKVPipeline,
silently downscaling any reference image above ~1MP. Expose it as a
max_area __call__ argument (default unchanged), mirroring the existing
max_area parameter of FluxKontextPipeline.

Author: sqhuang

src/diffusers/pipelines/flux2/pipeline_flux2.py

@@ -765,6 +765,7 @@ def __call__(
         max_sequence_length: int = 512,
         text_encoder_out_layers: tuple[int] = (10, 20, 30),
         caption_upsample_temperature: float = None,
+        max_area: int = 1024**2,
     ):
         r"""
         Function invoked when calling the pipeline for generation.
@@ -832,6 +833,9 @@ def __call__(
             caption_upsample_temperature (`float`):
                 When specified, we will try to perform caption upsampling for potentially improved outputs. We
                 recommend setting it to 0.15 if caption upsampling is to be performed.
+            max_area (`int`, defaults to `1024 ** 2`):
+                The maximum area (in pixels) allowed for each condition image. Condition images whose area exceeds
+                this value are downscaled to fit it while preserving their aspect ratio.
 
         Examples:
 
@@ -891,8 +895,8 @@ def __call__(
             condition_images = []
             for img in image:
                 image_width, image_height = img.size
-                if image_width * image_height > 1024 * 1024:
-                    img = self.image_processor._resize_to_target_area(img, 1024 * 1024)
+                if image_width * image_height > max_area:
+                    img = self.image_processor._resize_to_target_area(img, max_area)
                     image_width, image_height = img.size
 
                 multiple_of = self.vae_scale_factor * 2

src/diffusers/pipelines/flux2/pipeline_flux2_klein.py

@@ -632,6 +632,7 @@ def __call__(
         callback_on_step_end_tensor_inputs: list[str] = ["latents"],
         max_sequence_length: int = 512,
         text_encoder_out_layers: tuple[int] = (9, 18, 27),
+        max_area: int = 1024**2,
     ):
         r"""
         Function invoked when calling the pipeline for generation.
@@ -700,6 +701,9 @@ def __call__(
             max_sequence_length (`int` defaults to 512): Maximum sequence length to use with the `prompt`.
             text_encoder_out_layers (`tuple[int]`):
                 Layer indices to use in the `text_encoder` to derive the final prompt embeddings.
+            max_area (`int`, defaults to `1024 ** 2`):
+                The maximum area (in pixels) allowed for each condition image. Condition images whose area exceeds
+                this value are downscaled to fit it while preserving their aspect ratio.
 
         Examples:
 
@@ -769,8 +773,8 @@ def __call__(
             condition_images = []
             for img in image:
                 image_width, image_height = img.size
-                if image_width * image_height > 1024 * 1024:
-                    img = self.image_processor._resize_to_target_area(img, 1024 * 1024)
+                if image_width * image_height > max_area:
+                    img = self.image_processor._resize_to_target_area(img, max_area)
                     image_width, image_height = img.size
 
                 multiple_of = self.vae_scale_factor * 2

src/diffusers/pipelines/flux2/pipeline_flux2_klein_kv.py

@@ -627,6 +627,7 @@ def __call__(
         callback_on_step_end_tensor_inputs: list[str] = ["latents"],
         max_sequence_length: int = 512,
         text_encoder_out_layers: tuple[int] = (9, 18, 27),
+        max_area: int = 1024**2,
     ):
         r"""
         Function invoked when calling the pipeline for generation.
@@ -668,6 +669,9 @@ def __call__(
                 Maximum sequence length for the prompt.
             text_encoder_out_layers (`tuple[int]`):
                 Layer indices for text encoder hidden state extraction.
+            max_area (`int`, defaults to `1024 ** 2`):
+                The maximum area (in pixels) allowed for each condition image. Condition images whose area exceeds
+                this value are downscaled to fit it while preserving their aspect ratio.
 
         Examples:
 
@@ -720,8 +724,8 @@ def __call__(
             condition_images = []
             for img in image:
                 image_width, image_height = img.size
-                if image_width * image_height > 1024 * 1024:
-                    img = self.image_processor._resize_to_target_area(img, 1024 * 1024)
+                if image_width * image_height > max_area:
+                    img = self.image_processor._resize_to_target_area(img, max_area)
                     image_width, image_height = img.size
 
                 multiple_of = self.vae_scale_factor * 2