Update README.md

README.md

@@ -143,29 +143,19 @@ first three improvements and the first drop in a random draw from those images.
 | COCO 300233 | mixing bowl corn food | Ingredients and a food processor are arranged on a gray surface. (0.40) | **Ingredients and a food processor are laid out on a table. (0.47)** |
 | COCO 419228 (failure) | ice bear polar | **A white polar bear walks on rocks in an enclosure. (2.54)** | A large white polar bear walks on a rocky surface. (1.85) |
 
-**Summary: gain over each captioner's own top beam (CIDEr).**
+<!-- **Summary: gain over each captioner's own top beam (CIDEr).**
 
 | Captioner | Top beam | + This model | Gain [95% CI] |
 | --- | ---: | ---: | --- |
 | BLIP-base (training captioner) | 1.0632 | 1.1470 | +0.0838 [+0.073, +0.095] |
 | ViLBERT (paper's candidates) | 1.1150 | 1.1475 | +0.0325 [+0.021, +0.044] |
 | Gemma 4 12B | 0.9624 | 0.9783 | +0.0159 [+0.009, +0.023] |
+ -->
 
-## Examples (ViLBERT candidates, CIDEr per caption)
-
-| Image | Best Beam | This model |
-| --- | --- | --- |
-| COCO 460621 | a white toilet sitting next to a white bath tub (0.39) | **a bathroom with a sink and a toilet (0.83)** |
-| COCO 43680 | a group of cows that are standing in the dirt (0.64) | **a herd of cattle walking across a dry field (1.32)** |
-| COCO 424432 | a white and brown dog laying on a bed (0.73) | **a dog laying on a bed with a blanket (0.86)** |
-| COCO 365685 (failure) | **a woman is taking a bite out of a hot dog (0.92)** | a woman sitting at a table eating a hotdog (0.11) |
-
-When the model changes the caption (1,948 of 5,000 images), the change improves
-CIDEr 1,110 times and lowers it 838 times.
 
 ## Limitations
 
-- **Use it as a reranker, not a relatedness classifier.** Reward training shifts
+- **Use it as a reranker, not a (relatedness classifier)[https://huggingface.co/AhmedSSabir/ModernBERT-CNN-Caption-Reranker]. ** Reward training shifts
   the scores upward. On the paper's 41,849 held-out relatedness pairs, accuracy at
   the 0.5 cutoff falls from 0.893 to 0.631, and ROC-AUC from 0.962 to 0.937. For
   relatedness probabilities, use the stage I BERT-CNN.