feat/initial_implementation

README.md

@@ -10,7 +10,7 @@ The `FlashRankMultipleChoiceSolver` plugin is designed for the Open Voice OS (OV
 
 ### Important Note on FlashRank and Llama-CPP Compatibility
 
-Installing FlashRank can lead to a downgrade of the `llama-cpp-python` version, which is critical for GPU support and performance, especially for [ovos-solver-gguf-plugin](https://github.com/TigreGotico/ovos-solver-gguf-plugin). 
+Installing FlashRank can lead to a downgrade of the `llama-cpp-python` version, which is critical for GPU support and performance, especially for large language models (LLMs). This issue is tracked in [FlashRank's GitHub repository](https://github.com/PrithivirajDamodaran/FlashRank/issues/29).
 
 **Workaround for GPU Support with `llama-cpp-python`:**
 
@@ -19,7 +19,7 @@ If you need GPU support with `llama-cpp-python`, you might need to reinstall it
 CMAKE_ARGS="-DGGML_CUDA=on" FORCE_CMAKE=1 pip install llama-cpp-python --force-reinstall --no-cache-dir
 ```
 
-Be aware that installing FlashRank may undo these custom installations. This issue is tracked in [FlashRank's GitHub repository](https://github.com/PrithivirajDamodaran/FlashRank/issues/29).
+Be aware that installing FlashRank may undo these custom installations
 
 ## Usage
 
@@ -60,11 +60,15 @@ Example configuration:
 
 The following models are available for use with the `FlashRankMultipleChoiceSolver`:
 
+## Available Models
+
+The following models are available for use with the `FlashRankMultipleChoiceSolver`:
+
 | Model Name                   | Description                                                                                                   |
 |------------------------------|---------------------------------------------------------------------------------------------------------------|
 | ms-marco-TinyBERT-L-2-v2     | (default) [Model card](https://www.modelcards.com/ms-marco-TinyBERT-L-2-v2)                                     |
 | ms-marco-MiniLM-L-12-v2      | [Model card](https://www.modelcards.com/ms-marco-MiniLM-L-12-v2)                                               |
 | rank-T5-flan                 | Best non cross-encoder reranker [Model card](https://www.modelcards.com/rank-T5-flan)                          |
 | ms-marco-MultiBERT-L-12      | Multi-lingual, supports 100+ languages                                                                         |
 | ce-esci-MiniLM-L12-v2        | FT on Amazon ESCI dataset (This is interesting because most models are FT on MSFT MARCO Bing queries) [Model card](https://www.modelcards.com/ce-esci-MiniLM-L12-v2) |
-| rank_zephyr_7b_v1_full       | 4-bit-quantised GGUF [Model card](https://www.modelcards.com/rank_zephyr_7b_v1_full) (Offers very competitive performance, with large context window and relatively faster for a 4GB model) |
+| rank_zephyr_7b_v1_full       | 4-bit-quantised GGUF [Model card](https://www.modelcards.com/rank_zephyr_7b_v1_full) (Offers very competitive performance, with large context window and relatively faster for a 4GB model) |