Add Installation Option to Depend Only on ONNX, Excluding New Torch and CUDA Packages

Nobody has claimed this yet.

### Feature request

An installation option that allows users to install the project with dependencies limited to ONNX only, excluding newer versions of Torch and CUDA, particularly CUDA 12. This option would enable users to run the project in environments with older CUDA versions (e.g., 11.7), pure CPU setups, or other ONNX-compatible platforms like ONNX-DirectML.

There are some polular enbedding and reranker model has onnx binary like [youdao/bce-embedding](https://www.modelscope.cn/models/netease-youdao/bce-embedding-base_v1/files), and jina-embeddings-v2-base-zh.

### Motivation

Current installation requirements can create barriers for users with legacy hardware or specific configurations. By providing an option to install without the latest Torch and CUDA dependencies, we can enhance accessibility and flexibility, allowing more users to effectively utilize the project in a broader range of environments.

### Your contribution

Help test or fix few simple bugs.

Start by locating the project's installation requirements and the declarations that add Torch and CUDA packages. Compare the existing installation paths with the requested ONNX-only, CPU, and older-CUDA environments. Done means a documented installation option excludes the newer Torch and CUDA dependencies while supporting the stated ONNX-compatible setups.

Written by the indexing model from the issue text.