Add support for Embedding Tasks

### Issue 1
## Summary
> Wrap LiteLLM to support both Text and Image Embedding Models.

## Pain Points
> What challenges are users encountering without this feature?

1. Users may be limited to handling only one type of embedding model, either text or image, which restricts the versatility of their applications.
2. The lack of unified support for both model types may lead to inefficient workflows and integration issues.

## Current Workarounds
> Have you come across any alternatives or tried any workarounds?

1. Users might have to implement separate systems or tools for text and image embeddings, which can be complex and cumbersome.
2. Custom code might be used to bridge the gap between text and image embeddings, but this could result in inconsistent performance.

## Solution
> What solution would you like to see?

1. Wrap LiteLLM to provide seamless support for both text and image embedding models, simplifying integration and enhancing versatility.

## References
> Have you seen similar solutions or examples elsewhere?

1. Look into existing frameworks or libraries that offer support for both text and image embeddings for best practices and implementation guidance.

## Additional Notes
> Is there anything else you’d like us to know?

1. Integrating support for both embedding types will improve the functionality and usability of LiteLLM.

## Contribution
- [ ] I’d love help with this if needed!

該 issue 提到封裝 LiteLLM 以支援文字和圖像嵌入模型。首先，檢查程式碼庫中現有的 LiteLLM 整合，可能在 `cyyrus/llm/` 或 `cyyrus/integrations/` 等模組中。查看目前如何處理文字嵌入。然後，研究 LiteLLM 的嵌入 API 及其圖像嵌入能力。「完成」意味著包裝器可以接受文字和圖像輸入並返回嵌入，並已加入測試。

由索引模型根據 Issue 內容生成。