This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
SAVRN Model Hub · Models by Task
Sentence Similarity Models
73 open-weight sentence similarity models in the SAVRN Model Hub, with Sentence Transformers, Dell Research Harvard and Liang Wang publishing the most.
73Models
25Publishers
12M to 8.1BParameter range
5Licenses
SAVRN's Take
An embedding model turns a sentence or paragraph into a vector, and the rest is arithmetic on those vectors: semantic search, clustering, the retrieval step ahead of a language model. The 73 models here run from 12M parameters in paraphrase-albert-small-v2 to 8.1B in Qwen3-VL-Embedding-8B, and the 23M all-MiniLM-L6-v2 takes 0.1 GB at 16-bit on the Index's cheapest host, one MI300X at $1.85 an hour. Only the Qwen entry, which also embeds images, screenshots and video, needs real memory: 19.5 GB at 16-bit, 4.9 GB at 4-bit. A card already running generation has room for one of these beside it; the cost is in the text pushed through, not the weights.
Sentence Transformers and Dell Research Harvard each publish 17 of the 73, then Liang Wang with 6. Downloads concentrate harder: all-MiniLM-L6-v2 pulled 255,050,544 in a month, more than five times the 45,730,151 for paraphrase-multilingual-MiniLM-L12-v2, with bge-m3 third at 38,175,398. Licenses split 33 Apache-2.0, 12 MIT, one Gemma, one other, and 26 not stated. Settle that last group first; replacing an embedding model later means re-embedding the corpus.
Two figures matter more than parameter count. Context length: the Sentence Transformers and e5 models stop at 512 or 514 tokens, so documents get chunked, while nomic-embed-text-v1.5 takes 2,048, bge-m3 8,194 and the Qwen model 262,144. Vector width: the 384-dimensional output of all-MiniLM-L6-v2 stores at half the size of the 768-dimensional vectors from all-mpnet-base-v2, paid on every row. Check language coverage too, since multilingual-e5-small handles 100 languages, and whether the model wants a task prefix on every string, as nomic-embed-text-v1.5 does.
Most Downloaded
| Model | Publisher | Parameters | License | Monthly downloads | Cheapest GPUs at 16-bit |
|---|---|---|---|---|---|
| all-MiniLM-L6-v2 | Sentence Transformers | 23M | apache-2.0 | 255.1M | 1x MI300X, $1.85/hr |
| paraphrase-multilingual-MiniLM-L12-v2 | Sentence Transformers | 118M | apache-2.0 | 45.7M | 1x MI300X, $1.85/hr |
| bge-m3 | Beijing Academy of Artificial Intelligence | — | mit | 38.2M | — |
| all-mpnet-base-v2 | Sentence Transformers | 109M | apache-2.0 | 22.9M | 1x MI300X, $1.85/hr |
| nomic-embed-text-v1.5 | Nomic AI | 137M | apache-2.0 | 14.9M | 1x MI300X, $1.85/hr |
| multilingual-e5-small | Liang Wang | 118M | mit | 12.3M | 1x MI300X, $1.85/hr |
| paraphrase-multilingual-mpnet-base-v2 | Sentence Transformers | 278M | apache-2.0 | 9.8M | 1x MI300X, $1.85/hr |
| multilingual-e5-base | Liang Wang | 278M | mit | 7.6M | 1x MI300X, $1.85/hr |
| bge-base-en-v1.5-course-recommender-v5 | Data Society | 109M | Not stated | 4.2M | 1x MI300X, $1.85/hr |
| all-MiniLM-L12-v2 | Sentence Transformers | 33M | apache-2.0 | 4.2M | 1x MI300X, $1.85/hr |
Licenses
| License | Models | Commercial use |
|---|---|---|
| apache-2.0 | 33 | Yes |
| not stated | 26 | Not stated |
| mit | 12 | Yes |
| other | 1 | Read the license |
| gemma | 1 | Yes, with conditions |
Who Publishes Them
| Publisher | Models |
|---|---|
| Sentence Transformers | 17 |
| Dell Research Harvard | 17 |
| Liang Wang | 6 |
| Qdrant | 4 |
| Nomic AI | 3 |
| Alibaba-NLP | 3 |
All 73 Models, Page 2 of 2
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model - it just wraps around the class. This model has been fine-tuned on the model: sentence-transformers/all-mpnet-base-v2. It is pretrained for the language: en. To use this model using sentence-transformers: To use this model for clustering with LinkTransformer installed: We evaluate using the standard LinkTransformer information retrieval metrics. Our test set evaluations are available here. The model was trained with the parameters: torch.utils.data.dataloader.DataLoader of length 31 with parameters: linktransformer.modifiedsbert.losses.SupConLosswandb If you use our…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…
Semantic-Lite-2 is a lightweight multilingual sentence embedding model that produces 256-dimensional semantic vectors. It is designed for semantic search, sentence similarity, clustering, retrieval, and retrieval-augmented generation (RAG) tasks. The model is built on top of the Spark-X2.5-1.7B backbone using a frozen-backbone plus trainable-projection-head approach. The 256-dimensional output keeps vector storage compact while preserving strong retrieval quality. Vectors are L2-normalized, so cosine similarity is computed as a simple dot product. Evaluated on 500 Indonesian NLI evaluation pairs (retrieval task, chance level 0.2%): Cross-lingual evaluation (10 languages, 20 pairs per…
Questions
Which Sentence similarity models are most downloaded?
By monthly downloads reported by the Hugging Face Hub: lt-un-data-fine-industry-en (79); lt-wikidata-comp-es (79); lt-un-data-fine-coarse-es (78).