Sentence Similarity
sentence-transformers
Safetensors
bert
information-retrieval
biomedical
metasyn
text-embeddings-inference
Instructions to use BFTree/MA-Retriever with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use BFTree/MA-Retriever with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("BFTree/MA-Retriever") 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
| { | |
| "base_model": "BAAI/bge-large-en-v1.5", | |
| "base_model_revision": "d4aa6901d3a41ba39fb536a557fa166f842b0e09", | |
| "seed": 718, | |
| "query_template": "Research_Question + PI/ECO with BGE retrieval instruction", | |
| "document_template": "Title + Abstract", | |
| "train_reviews": 336, | |
| "test_reviews": 86, | |
| "model_selection_reviews": 40, | |
| "model_selection_rule": "Recall@20, then Recall@100, then Recall@200", | |
| "selected_epochs": 1, | |
| "final_retraining_reviews": 336, | |
| "training_reviews_with_positive_pairs": 334, | |
| "test_linked_articles_excluded_from_positive_training": 1649, | |
| "constructed_training_pairs": 5585, | |
| "complete_batch_examples_used": 5568, | |
| "batch_size": 64, | |
| "learning_rate": 2e-05, | |
| "max_seq_length": 512, | |
| "loss": "MultipleNegativesRankingLoss", | |
| "batch_sampler": "NoDuplicatesDataLoader", | |
| "epochs": 1, | |
| "test_metrics": { | |
| "recall@5": 0.24032354463558983, | |
| "recall@10": 0.38838341990475195, | |
| "recall@20": 0.534827514606953, | |
| "recall@50": 0.7526966587841311, | |
| "recall@100": 0.8420887384792756, | |
| "recall@200": 0.917380895710928, | |
| "precision@5": 0.4790697674418605, | |
| "precision@10": 0.4197674418604652, | |
| "precision@20": 0.3401162790697674, | |
| "precision@50": 0.22604651162790698, | |
| "precision@100": 0.14046511627906977, | |
| "precision@200": 0.08406976744186047 | |
| } | |
| } | |