Update README.md
Browse files
README.md
CHANGED
|
@@ -40,3 +40,34 @@ The models were evaluated using popular metrics for image captioning: **BLEU (1-
|
|
| 40 |
| **ViT + GPT2** | **0.728** | 0.545 | 0.385 | 0.265 | **0.502** | 0.532 |
|
| 41 |
|
| 42 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
| **ViT + GPT2** | **0.728** | 0.545 | 0.385 | 0.265 | **0.502** | 0.532 |
|
| 41 |
|
| 42 |
---
|
| 43 |
+
|
| 44 |
+
## **Inference Example**
|
| 45 |
+
|
| 46 |
+
Below is an example of how the models perform on a given image. The table shows the reference caption and the predicted captions generated by each model.
|
| 47 |
+
|
| 48 |
+
<table>
|
| 49 |
+
<tr>
|
| 50 |
+
<th>Image</th>
|
| 51 |
+
<th>Reference Caption</th>
|
| 52 |
+
<th>Predicted Caption</th>
|
| 53 |
+
</tr>
|
| 54 |
+
<tr>
|
| 55 |
+
<td>
|
| 56 |
+
<img src="examples/000000166391.jpg" alt="Traffic light" width="300">
|
| 57 |
+
</td>
|
| 58 |
+
<td>
|
| 59 |
+
<ol>
|
| 60 |
+
<li>Traffic is stopped at a red stop light.</li>
|
| 61 |
+
<li>Cars are stopped at a traffic light on a highway.</li>
|
| 62 |
+
<li>A number of red and green traffic lights on a wide highway.</li>
|
| 63 |
+
<li>A large and wide street covered in lots of traffic lights.</li>
|
| 64 |
+
<li>A traffic light and intersection with cars traveling in both directions on the street.</li>
|
| 65 |
+
</ol>
|
| 66 |
+
</td>
|
| 67 |
+
<td>
|
| 68 |
+
<b>ResNet50 + LSTM:</b> a traffic light with a street sign on it.<br>
|
| 69 |
+
<b>ViT + BERT:</b> a bunch of traffic lights hanging from a wire.<br>
|
| 70 |
+
<b>ViT + GPT2:</b> A green traffic light hanging over a street.
|
| 71 |
+
</td>
|
| 72 |
+
</tr>
|
| 73 |
+
</table>
|