Update README.md
Browse files
README.md
CHANGED
|
@@ -17,6 +17,8 @@ The model produces **384-dimensional** embeddings (CLS token) optimized for cosi
|
|
| 17 |
|
| 18 |
# Training details
|
| 19 |
|
|
|
|
|
|
|
| 20 |
* Max sequence length: 128 tokens
|
| 21 |
|
| 22 |
* Loss: MultipleNegativesSymmetricRankingLoss (InfoNCE)
|
|
@@ -37,4 +39,4 @@ On a small test set, the model achieves:
|
|
| 37 |
|
| 38 |
* Margin: 0.44
|
| 39 |
|
| 40 |
-
#
|
|
|
|
| 17 |
|
| 18 |
# Training details
|
| 19 |
|
| 20 |
+
* ~38.3M parameters
|
| 21 |
+
|
| 22 |
* Max sequence length: 128 tokens
|
| 23 |
|
| 24 |
* Loss: MultipleNegativesSymmetricRankingLoss (InfoNCE)
|
|
|
|
| 39 |
|
| 40 |
* Margin: 0.44
|
| 41 |
|
| 42 |
+
# Usage is coming soon!
|