infinity
infinity copied to clipboard
michaelfeil
Infinity is a high-throughput, low-latency REST API for serving vector embeddings, supporting a wide range of text-embedding models and frameworks.
## Related Issue The problem stems from the `__init__` method of the parent class, `SentenceTransformer`. This method uses `self.__class__.__name__` to [set a configuration value](https://github.com/UKPLab/sentence-transformers/blob/d5c8f5181daa468c63041fde18c659c4e4267e77/sentence_transformers/SentenceTransformer.py#L200). When subclass `SentenceTransformerPatched` calls `super().__init__()`, `self.__class__.__name__`...
numpy 1 do not support python >= 3.13 and from [link](https://github.com/microsoft/onnxruntime/issues/21063#issuecomment-2389415124) onnxruntime>=1.19.0 (released August 2024) is compatible with numpy>=2.0.0. and we have already pin onnxruntime to 1.19. So I think...
### Feature request The latest Qwen3 embedding and rerank models suggest use of user-defined instructions for better performance, do we have plan to add support for this? Thanks. ### Motivation...
Metadata
Owner
Metadata
Infinity is a high-throughput, low-latency REST API for serving vector embeddings, supporting a wide range of text-embedding models and frameworks.