Commit
·
d81c518
1
Parent(s):
d97ecc6
Update README.md
Browse files
README.md
CHANGED
|
@@ -53,6 +53,26 @@ We are forking sentence-transformers/all-MiniLM-L6-v2 as it is similar to the ta
|
|
| 53 |
- Used the AdamW optimizer with a 2e-5 learning rate.
|
| 54 |
- The full training script is accessible in this current repository: `train_script.py`.
|
| 55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
#### Datasets
|
| 57 |
| Dataset | Paper | Number of training tuples |
|
| 58 |
|--------------------------------------------------------|:----------------------------------------:|:--------------------------:|
|
|
|
|
| 53 |
- Used the AdamW optimizer with a 2e-5 learning rate.
|
| 54 |
- The full training script is accessible in this current repository: `train_script.py`.
|
| 55 |
|
| 56 |
+
|
| 57 |
+
#### Performance
|
| 58 |
+
|
| 59 |
+
| Model Name | Performance Sentence Embeddings (14 Datasets) | Performance Semantic Search (6 Datasets) | Avg. Performance | Speed | Model Size |
|
| 60 |
+
|----------------------------------------------|-----------------------------------------------|----------------------------------------|-----------------|-------|------------|
|
| 61 |
+
| all-mpnet-base-v2 | 69.57 | 57.02 | 63.30 | 2800 | 420 MB |
|
| 62 |
+
| multi-qa-mpnet-base-dot-v1 | 66.76 | 57.60 | 62.18 | 2800 | 420 MB |
|
| 63 |
+
| all-distilroberta-v1 | 68.73 | 50.94 | 59.84 | 4000 | 290 MB |
|
| 64 |
+
| all-MiniLM-L12-v2 | 68.70 | 50.82 | 59.76 | 7500 | 120 MB |
|
| 65 |
+
| multi-qa-distilbert-cos-v1 | 65.98 | 52.83 | 59.41 | 4000 | 250 MB |
|
| 66 |
+
| **all-MiniLM-L6-v2** | **68.06** | **49.54** | **58.80** | **14200** | **80 MB** |
|
| 67 |
+
| multi-qa-MiniLM-L6-cos-v1 | 64.33 | 51.83 | 58.08 | 14200 | 80 MB |
|
| 68 |
+
| paraphrase-multilingual-mpnet-base-v2 | 65.83 | 41.68 | 53.75 | 2500 | 970 MB |
|
| 69 |
+
| paraphrase-albert-small-v2 | 64.46 | 40.04 | 52.25 | 5000 | 43 MB |
|
| 70 |
+
| paraphrase-multilingual-MiniLM-L12-v2 | 64.25 | 39.19 | 51.72 | 7500 | 420 MB |
|
| 71 |
+
| paraphrase-MiniLM-L3-v2 | 62.29 | 39.19 | 50.74 | 19000 | 61 MB |
|
| 72 |
+
| distiluse-base-multilingual-cased-v1 | 61.30 | 29.87 | 45.59 | 4000 | 480 MB |
|
| 73 |
+
| distiluse-base-multilingual-cased-v2 | 60.18 | 27.35 | 43.77 | 4000 | 480 MB |
|
| 74 |
+
|
| 75 |
+
|
| 76 |
#### Datasets
|
| 77 |
| Dataset | Paper | Number of training tuples |
|
| 78 |
|--------------------------------------------------------|:----------------------------------------:|:--------------------------:|
|