Feature Extraction
ONNX
sentence-transformers
multilingual
onnxruntime
bert
e5
int8
text-embeddings-inference
Instructions to use orca-zhang/multilingual-e5-small-int8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use orca-zhang/multilingual-e5-small-int8 with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("orca-zhang/multilingual-e5-small-int8") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
multilingual-e5-small-int8 for Manticore
This repository is the immutable ONNX INT8 bundle used by ZimaOS Photos with
Manticore automatic embeddings. The root model.onnx is copied from
Xenova/multilingual-e5-small at revision
761b726dd34fb83930e26aab4e9ac3899aa1fa78, artifact
onnx/model_int8.onnx.
The embedding contract is fixed:
- output dimension: 384;
- query text prefix:
query:; - indexed caption prefix:
passage:; - pooling: attention-mask mean pooling followed by L2 normalization;
- maximum sequence length: 512 tokens.
ZimaOS Photos stores bilingual VLM captions and detailed summaries as passages.
Manticore downloads config.json, tokenizer.json, and the single root
model.onnx automatically. No safetensors, GGUF, or alternate ONNX weights are
published in this repository.
See PROVENANCE.json, 384D-prefix-fidelity-report.json, and SHA256SUMS for
the pinned sources and validation results.
- Downloads last month
- 807