AnglE-optimized Text Embeddings
Paper • 2309.12871 • Published • 3
Feature Extraction
Transformers
Safetensors
bert
mteb
Eval Results (legacy)
text-embeddings-inference
# Load model directly
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("technicolor/Angle_BERT")
model = AutoModel.from_pretrained("technicolor/Angle_BERT")Quick Links
Embedding model trained on google-bert/bert-base-uncased with the AnglE-optimized Text Embeddings
- Downloads last month
- 144
Spaces using technicolor/Angle_BERT 11
Paper for technicolor/Angle_BERT
Evaluation results
- accuracy on MTEB AmazonCounterfactualClassification (en)test set self-reported77.940
- ap on MTEB AmazonCounterfactualClassification (en)test set self-reported41.462
- f1 on MTEB AmazonCounterfactualClassification (en)test set self-reported71.913
- accuracy on MTEB AmazonPolarityClassificationtest set self-reported75.999
- ap on MTEB AmazonPolarityClassificationtest set self-reported70.686
- f1 on MTEB AmazonPolarityClassificationtest set self-reported75.804
- accuracy on MTEB AmazonReviewsClassification (en)test set self-reported37.184
- f1 on MTEB AmazonReviewsClassification (en)test set self-reported36.928
- map_at_1 on MTEB ArguAnatest set self-reported18.208
- map_at_10 on MTEB ArguAnatest set self-reported31.217
- map_at_100 on MTEB ArguAnatest set self-reported32.504
- map_at_1000 on MTEB ArguAnatest set self-reported32.543
- map_at_20 on MTEB ArguAnatest set self-reported32.048
- map_at_3 on MTEB ArguAnatest set self-reported26.790
- map_at_5 on MTEB ArguAnatest set self-reported29.176
- mrr_at_1 on MTEB ArguAnatest set self-reported18.990
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="technicolor/Angle_BERT")