issues
search
juanmc2005
/
SimilarityLearning
Similarity Learning applied to Speaker Verification and Semantic Textual Similarity
12
stars
7
forks
source link
Newest Changes
#29
Closed
juanmc2005
closed
5 years ago
juanmc2005
commented
5 years ago
Changelog
Custom seed, margin and thershold are now script parameters
SemEval modes are now augmentation objects
Separated SemEval evaluation experiments to cover logit and embedding based evaluations
t-SNE visualization is possible in STS, while still tracking the original phrases for each embedding
SemEval evaluators refactored into class hierarchy
Experiments have been splitted into different files
Added customized experience id for each script launch, which changes the name of the output models and logs
Mean loss for each epoch is now being logged to a file too
Added RegularModelSaver to save models every x epochs
Removed mutable default parameters
Added STS forward mode objects to change the way sentences are fed to the model (single, pair, triplets)
Bugfixes and minor improvements
Changelog