Sentence Similarity
sentence-transformers
Safetensors
bert
intent-classification
multilingual
layer-pruning
vocab-pruning
text-embeddings-inference
Instructions to use gomyk/intent-student-L3_compact with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use gomyk/intent-student-L3_compact with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("gomyk/intent-student-L3_compact") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
L3_compact
Ultra-compact multilingual sentence encoder (~50.9MB) for intent classification. 3 layers [0,6,11] + 20K vocab
Performance
| Model | Size | MassiveIntent | MassiveScenario | Average |
|---|---|---|---|---|
| Teacher (12L, full) | ~480MB | 55.52% | 61.01% | 58.27% |
| L6_bottom (38K vocab) | 98MB | 54.70% | 59.39% | 57.05% |
| L3_compact | 50.9MB | 47.5% | 50.4% | 48.95% |
Model Details
| Property | Value |
|---|---|
| Teacher | paraphrase-multilingual-MiniLM-L12-v2 |
| Vocab | ~20,000 (frequency-based pruning, 97.4% coverage) |
| Size | 50.9MB |
| Distilled | No |
Quick Start
- Downloads last month
- 3