Update README.md
Browse files
README.md
CHANGED
|
@@ -79,12 +79,6 @@ There are two options to calculate semantic compatibility between an image and a
|
|
| 79 |
|
| 80 |
### Cosine Similarity
|
| 81 |
|
| 82 |
-
```python
|
| 83 |
-
import torch.nn.functional as F
|
| 84 |
-
|
| 85 |
-
similarity = F.cosine_similarity(image_embedding, text_embedding)
|
| 86 |
-
```
|
| 87 |
-
|
| 88 |
The `similarity` will belong to the `[-1, 1]` range, `1` meaning the absolute match.
|
| 89 |
|
| 90 |
__Pros__:
|
|
@@ -103,10 +97,6 @@ __Cons__:
|
|
| 103 |
Unlike cosine similarity, unimodal embedding are not enough.
|
| 104 |
Joint embedding will be needed and the resulting `score` will belong to the `[0, 1]` range, `1` meaning the absolute match.
|
| 105 |
|
| 106 |
-
```python
|
| 107 |
-
score = model.get_matching_scores(joint_embedding)
|
| 108 |
-
```
|
| 109 |
-
|
| 110 |
__Pros__:
|
| 111 |
|
| 112 |
- Joint embedding captures fine-grained features.
|
|
|
|
| 79 |
|
| 80 |
### Cosine Similarity
|
| 81 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 82 |
The `similarity` will belong to the `[-1, 1]` range, `1` meaning the absolute match.
|
| 83 |
|
| 84 |
__Pros__:
|
|
|
|
| 97 |
Unlike cosine similarity, unimodal embedding are not enough.
|
| 98 |
Joint embedding will be needed and the resulting `score` will belong to the `[0, 1]` range, `1` meaning the absolute match.
|
| 99 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
__Pros__:
|
| 101 |
|
| 102 |
- Joint embedding captures fine-grained features.
|