Model card: task collage, editing and generation prompts, LoRA adapters

@@ -46,7 +46,9 @@ Install PyTorch with CUDA support first, then:
 pip install "diffusers>=0.37.0" transformers accelerate safetensors pillow
 ```
-The fine-tuned transformer is swapped into the [FLUX.2 [klein] base 9B](https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B) pipeline. One model handles every task and the prompt selects it. This example runs <a href="#quickstart"><img src="assets/icons/semseg.png" width="26" alt=""></a> semantic segmentation:
 ```python
 import torch
@@ -92,7 +94,7 @@ output.resize(image.size, Image.NEAREST).save("output.png")
 Replace `prompt` above with one of these.
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/refseg.png" width="26" alt=""></a> **Referring segmentation**
 ```python
 prompt = (
@@ -102,7 +104,7 @@ prompt = (
 )
 ```
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/instanceseg.png" width="26" alt=""></a> **Instance segmentation**
 ```python
 prompt = (
@@ -120,7 +122,7 @@ prompt = (
 )
 ```
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/depth.png" width="26" alt=""></a> **Depth estimation**
 ```python
 prompt = (
@@ -133,7 +135,7 @@ prompt = (
 )
 ```
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/normals.png" width="26" alt=""></a> **Surface normal estimation**
 ```python
 prompt = (
@@ -144,7 +146,7 @@ prompt = (
 )
 ```
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/imgedit.png" width="26" alt=""></a> **Image editing**
 ```python
 prompt = (
@@ -152,7 +154,7 @@ prompt = (
 )
 ```
-<a href="#quickstart-prompts-for-the-other-tasks"><img src="assets/icons/generation.png" width="26" alt=""></a> **Text-to-image generation**
 Call the pipeline without `image` and give it a size:

 pip install "diffusers>=0.37.0" transformers accelerate safetensors pillow
 ```
+The fine-tuned transformer is swapped into the [FLUX.2 [klein] base 9B](https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B) pipeline. One model handles every task and the prompt selects it. This example runs:
+<img src="assets/labels/semseg.png" width="255" alt="Semantic segmentation">
 ```python
 import torch
 Replace `prompt` above with one of these.
+<img src="assets/labels/refseg.png" width="257" alt="Referring segmentation">
 ```python
 prompt = (
 )
 ```
+<img src="assets/labels/instanceseg.png" width="248" alt="Instance segmentation">
 ```python
 prompt = (
 )
 ```
+<img src="assets/labels/depth.png" width="194" alt="Depth estimation">
 ```python
 prompt = (
 )
 ```
+<img src="assets/labels/normals.png" width="282" alt="Surface normal estimation">
 ```python
 prompt = (
 )
 ```
+<img src="assets/labels/imgedit.png" width="161" alt="Image editing">
 ```python
 prompt = (
 )
 ```
+<img src="assets/labels/generation.png" width="271" alt="Text-to-image generation">
 Call the pipeline without `image` and give it a size: