Instructions to use IngridCEK/simcse-bert-base-snli-sup with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use IngridCEK/simcse-bert-base-snli-sup with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("IngridCEK/simcse-bert-base-snli-sup") sentences = [ "That is a happy person", "That is a happy dog", "That is a very happy person", "Today is a sunny day" ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [4, 4] - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -22,8 +22,6 @@ max_len train 32 / eval 128 · AdamW (wd 0.01) · warmup 0.1 lineal · semilla 4
|
|
| 22 |
|---|---|---|---|
|
| 23 |
| 82.41 | 78.06 | 0.195 | -3.079 |
|
| 24 |
|
| 25 |
-
## Limitaciones
|
| 26 |
-
Entrenado con ~33k pares de un solo dominio (SNLI, descripciones de imágenes): el rendimiento fuera de ese dominio no se evaluó. Solo inglés, oraciones cortas (≤128 tokens). El rendimiento queda por debajo del paper porque se usaron menos datos y otros hiperparámetros. No evalúa negación ni antónimos de forma confiable.
|
| 27 |
|
| 28 |
## Uso
|
| 29 |
```python
|
|
|
|
| 22 |
|---|---|---|---|
|
| 23 |
| 82.41 | 78.06 | 0.195 | -3.079 |
|
| 24 |
|
|
|
|
|
|
|
| 25 |
|
| 26 |
## Uso
|
| 27 |
```python
|