LEMUR and Mean Centering for Late-Interaction Retrieval in txtai
• 9
AI & ML interests
Applying machine learning to solve everyday problems
Articles
Set of BERT models with a modified embeddings layer
Add knowledge to your txtai agents and processes.
LiteRT is Google's on-device framework for high-performance ML & GenAI deployment on edge platforms.
Models for working with medical and scientific literature.
-
NeuML/pubmedbert-base-embeddings
Sentence Similarity • 0.1B • Updated • 1.06M • • 207 -
NeuML/pubmedbert-base-embeddings-matryoshka
Sentence Similarity • 0.1B • Updated • 1.9k • • 23 -
NeuML/pubmedbert-base-embeddings-8M
Sentence Similarity • 7.81M • Updated • 384 • 10 -
NeuML/pubmedbert-base-colbert
Sentence Similarity • 0.1B • Updated • 89 • 7
Small models (22.7M) trained for specific domains
Embeddings indexes and datasets for Wikipedia data.
Late interaction models
StaticVectors models to detect language. Exports of FastText that run in NumPy without needing FastText
Datasets with medical and scientific literature.
Models distilled with Model2Vec - 100K / 500K / 1M / 2M / 8M parameter variants.
-
NeuML/pubmedbert-base-embeddings-8M
Sentence Similarity • 7.81M • Updated • 384 • 10 -
NeuML/pubmedbert-base-embeddings-2M
Sentence Similarity • 1.95M • Updated • 321 • 3 -
NeuML/pubmedbert-base-embeddings-1M
Sentence Similarity • 1M • Updated • 330 • 2 -
NeuML/pubmedbert-base-embeddings-500K
Sentence Similarity • 500k • Updated • 338 • 2
Text to Speech (TTS) models compatible with txtai's TextToSpeech pipeline.
Legacy word vectors (FastText, GloVe, Word2Vec) stored in the StaticVectors format
Set of BERT models with a modified embeddings layer
Late interaction models
Add knowledge to your txtai agents and processes.
StaticVectors models to detect language. Exports of FastText that run in NumPy without needing FastText
LiteRT is Google's on-device framework for high-performance ML & GenAI deployment on edge platforms.
Datasets with medical and scientific literature.
Models for working with medical and scientific literature.
-
NeuML/pubmedbert-base-embeddings
Sentence Similarity • 0.1B • Updated • 1.06M • • 207 -
NeuML/pubmedbert-base-embeddings-matryoshka
Sentence Similarity • 0.1B • Updated • 1.9k • • 23 -
NeuML/pubmedbert-base-embeddings-8M
Sentence Similarity • 7.81M • Updated • 384 • 10 -
NeuML/pubmedbert-base-colbert
Sentence Similarity • 0.1B • Updated • 89 • 7
Models distilled with Model2Vec - 100K / 500K / 1M / 2M / 8M parameter variants.
-
NeuML/pubmedbert-base-embeddings-8M
Sentence Similarity • 7.81M • Updated • 384 • 10 -
NeuML/pubmedbert-base-embeddings-2M
Sentence Similarity • 1.95M • Updated • 321 • 3 -
NeuML/pubmedbert-base-embeddings-1M
Sentence Similarity • 1M • Updated • 330 • 2 -
NeuML/pubmedbert-base-embeddings-500K
Sentence Similarity • 500k • Updated • 338 • 2
Small models (22.7M) trained for specific domains
Text to Speech (TTS) models compatible with txtai's TextToSpeech pipeline.
Embeddings indexes and datasets for Wikipedia data.
Legacy word vectors (FastText, GloVe, Word2Vec) stored in the StaticVectors format