v1.5.0

Notable Changes

• ONNX runtime for CPU deployments: greatly improve CPU deployment throughput
• Add /similarity route

What's Changed

• tokenizer max limit on input size by @ErikKaum in https://github.com/huggingface/text-embeddings-inference/pull/324
• docs: air-gapped deployments by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/326
• feat(onnx): add onnx runtime for better CPU perf by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/328
• feat: add /similarity route by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/331
• fix(ort): fix mean pooling by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/332
• chore(candle): update flash attn by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/335
• v1.5.0 by @OlivierDehaene in https://github.com/huggingface/text-embeddings-inference/pull/336

New Contributors

• @ErikKaum made their first contribution in https://github.com/huggingface/text-embeddings-inference/pull/324

Full Changelog: https://github.com/huggingface/text-embeddings-inference/compare/v1.4.0...v1.5.0