Update README.md

README.md

@@ -11,24 +11,9 @@ base_model: black-forest-labs/FLUX.1-Fill-dev
 ## Training Script:
 https://github.com/lrzjason/T2ITrainer
 
-# Object Removal LoRA of Flux Fill Dev v2.0
-
-## Dataset
-v2.0 \
-removal_timestep_alpha-2-1740.safetensors \
-trained with a subset of RORD Dataset \
-randomly selected with around 500 pairs \
-rank 16
-lr 1e-4
-epoch 3
-reg_timestep 700
-
-https://huggingface.co/datasets/lrzjason/ObjectRemovalAlpha \
-All images gather from unsplash. Thanks to all creators who created the awesome images.
-
 ## Model Description
 Update 20250517 \
-Released Remove Anything on runninghub. \
+Released Remove Anything v1.0 on runninghub. \
 https://www.runninghub.ai/ai-detail/1923389726278283265 \
 I could say much better than Object Removal LoRA v2 \
 and it is stronger than many closed source platform on object removal.
@@ -37,12 +22,27 @@ and it is stronger than many closed source platform on object removal.
     <tr>
       <td align="center">
         <p>Compare to closed source removal</p>
-        <img src="https://huggingface.co/lrzjason/ObjectRemovalFluxFill/resolve/main/ComfyUI_temp_npvyv_00007_.png" alt="Compare Between Alpha and Beta" width="800" />
+        <img src="https://huggingface.co/lrzjason/ObjectRemovalFluxFill/blob/main/ComfyUI_temp_ziejk_00002_.png" alt="Compare to closed source removal" width="1200" />
       </td>
     </tr>
   </table>
 </div>
 
+# Object Removal LoRA of Flux Fill Dev v2.0
+
+## Dataset
+v2.0 \
+removal_timestep_alpha-2-1740.safetensors \
+trained with a subset of RORD Dataset \
+randomly selected with around 500 pairs \
+rank 16
+lr 1e-4
+epoch 3
+reg_timestep 700
+
+https://huggingface.co/datasets/lrzjason/ObjectRemovalAlpha \
+All images gather from unsplash. Thanks to all creators who created the awesome images.
+
 Update 20250503 \
 Upload v2.0 checkpoint, removal_timestep_alpha-2-1740.safetensors. Trained on small subset to verify new training method, select objective with provided timestep, 700.