pplx-embed-v1-ron-32768

This model is a 23.29% smaller version of perplexity-ai/pplx-embed-v1-0.6b optimized for Romanian language via vocabulary size reduction using the trimming method.
This trimmed model should perform similarly to the original model with only 16,384 tokens and a much smaller memory footprint. However, it may not perform well for other languages as tokens not commonly used in the selected languages were removed from the vocabulary.

Model Statistics

Metric Original Trimmed Reduction
Vocabulary size 151,936 tokens 16,384 tokens 89.22%
Model size 596,049,920 params 457,244,672 params 23.29%

image

Mining Dataset Statistics

Usage

from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("alphaedge-ai/pplx-embed-v1-ron-16384")
# Run inference with queries and documents
query = "My query in Romanian"
documents = [
    "Chunk in Romanian",
    "Chunk in Romanian",
    "Chunk in Romanian",
]
query_embeddings = model.encode_query(query)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# Compute similarities to determine a ranking
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)

Citations

pplx-embed-v1

@misc{eslami2026diffusionpretraineddensecontextualembeddings,
      title={Diffusion-Pretrained Dense and Contextual Embeddings}, 
      author={Sedigheh Eslami and Maksim Gaiduk and Markus Krimmel and Louis Milliken and Bo Wang and Denis Bykov},
      year={2026},
      eprint={2602.11151},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2602.11151}, 
}

Trimming blog post

@misc{hf_blogpost_trimming,
      title={Introduction to Trimming}, 
      author={Loïck BOURDOIS and Tom AARSEN and Bram VANROY and Christopher AKIKI and Woojun JUNG and Manuel ROMERO and Prithiv SAKTHI},
      year={2026},
      url={https://huggingface.co/blog/lbourdois/introduction-to-trimming}, 
}
Downloads last month
17
Safetensors
Model size
0.5B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for alphaedge-ai/pplx-embed-v1-ron-16384

Quantized
(63)
this model

Dataset used to train alphaedge-ai/pplx-embed-v1-ron-16384

Collection including alphaedge-ai/pplx-embed-v1-ron-16384

Paper for alphaedge-ai/pplx-embed-v1-ron-16384