I train a BI-Encoder to get an Augmented SBERT and I get a final training result.
How can I interpret the following output of the final training result?
EmbeddingSimilarityEvaluator: Evaluating the model on test dataset:
Cosine-Similarity : Pearson: 0.8115 Spearman: 0.7777
Manhattan-Distance: Pearson: 0.7318 Spearman: 0.6822
Euclidean-Distance: Pearson: 0.7332 Spearman: 0.6835
Dot-Product-Similarity: Pearson: 0.7780 Spearman: 0.7543
0.7777387754875323 # output of test_evaluator(...)
The output result out of the following code snipped:
# load the stored augmented-sbert model
bi_encoder = SentenceTransformer(bi_encoder_path)
test_evaluator = EmbeddingSimilarityEvaluator.from_input_examples(test_samples, name='sts-test')
test_evaluator(bi_encoder, output_path=bi_encoder_path)
Is a high or low Person resp. Spearman better? They give information about the correlation.
