elastic/multilingual-e5-small-optimized

A quantized version of multilingual-e5-small. Quantization was performed per-layer under the same conditions as our ELSERv2 model, as described here.

Text Embeddings by Weakly-Supervised Contrastive Pre-training. Liang Wang, Nan Yang, Xiaolong Huang, Binxing Jiao, Linjun Yang, Daxin Jiang, Rangan Majumder, Furu Wei, arXiv 2022

Benchmarks

We performed a number of small benchmarks to assess both the changes in quality as well as inference latency against the baseline original model.

Quality

Measuring NDCG@10 using the dev split of the MIRACL datasets for select languages, we see mostly a marginal change in quality of the quantized model.

To test the English out-of-domain performance, we used the test split of various datasets in the BEIR evaluation. Measuring NDCG@10, we see a larger change in SCIFACT, but marginal in the other datasets evaluated.

Performance

Using a PyTorch model traced for Linux and Intel CPUs, we performed performance benchmarking with various lengths of input. Overall, we see on average a 50-20% performance improvement with the optimized model.

Disclaimer

This e5 model, as defined, hosted, integrated and used in conjunction with our other Elastic Software is covered by our standard warranty.