---
license: mit
language:
- en
tags:
- sparse sparsity quantized onnx embeddings int8
---
This is the quantized (INT8) ONNX variant of the [bge-large-en-v1.5](https://huggingface.co/BAAI/bge-large-en-v1.5) embeddings model created with [DeepSparse Optimum](https://github.com/neuralmagic/optimum-deepsparse) for ONNX export/inference pipeline and Neural Magic's [Sparsify](https://github.com/neuralmagic/sparsify) for one-shot quantization.

Model achieves 100% accuracy recovery on the STSB validation dataset vs. [dense ONNX variant](https://huggingface.co/zeroshot/bge-large-en-v1.5-dense).

Current list of sparse and quantized bge ONNX models:

| Links                                                                                               | Sparsification Method |
| --------------------------------------------------------------------------------------------------- | ---------------------- |
| [zeroshot/bge-large-en-v1.5-sparse](https://huggingface.co/zeroshot/bge-large-en-v1.5-sparse)     |    Quantization (INT8) & 50% Pruning                    |
| [zeroshot/bge-large-en-v1.5-quant](https://huggingface.co/zeroshot/bge-large-en-v1.5-quant)     |   Quantization (INT8)                     |
| [zeroshot/bge-base-en-v1.5-sparse](https://huggingface.co/zeroshot/bge-base-en-v1.5-sparse)     |   Quantization (INT8) & 50% Pruning                     |
| [zeroshot/bge-base-en-v1.5-quant](https://huggingface.co/zeroshot/bge-base-en-v1.5-quant)     |     Quantization (INT8)                    |
| [zeroshot/bge-small-en-v1.5-sparse](https://huggingface.co/zeroshot/bge-small-en-v1.5-sparse) |    Quantization (INT8) & 50% Pruning                    |
| [zeroshot/bge-small-en-v1.5-quant](https://huggingface.co/zeroshot/bge-small-en-v1.5-quant) |     Quantization (INT8)                    |

```bash
pip install -U deepsparse-nightly[sentence_transformers]
```

```python
from deepsparse.sentence_transformers import SentenceTransformer
model = SentenceTransformer('zeroshot/bge-large-en-v1.5-quant', export=False)

# Our sentences we like to encode
sentences = ['This framework generates embeddings for each input sentence',
    'Sentences are passed as a list of string.',
    'The quick brown fox jumps over the lazy dog.']

# Sentences are encoded by calling model.encode()
embeddings = model.encode(sentences)

# Print the embeddings
for sentence, embedding in zip(sentences, embeddings):
    print("Sentence:", sentence)
    print("Embedding:", embedding.shape)
    print("")
```

For further details regarding DeepSparse & Sentence Transformers integration, refer to the [DeepSparse README](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/sentence_transformers).


For general questions on these models and sparsification methods, reach out to the engineering team on our [community Slack](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ).

![;)](https://media.giphy.com/media/bYg33GbNbNIVzSrr84/giphy-downsized-large.gif)