RAG-Techniques

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #llm #machine-learning #rust #java #windows #awesome #vue #ios #mcp #open-source

(★ 10)

No description available.

RAG-Techniques Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# RAG-Techniques
---
This project aims at presenting common use-cases of RAG systems supported by codes. So far, the repo contains the following ***RAG*** techniques.
1. **[Adaptive Retrieval](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/adaptive_retrieval.py)**: A retrieval method that dynamically adjusts its strategies or parameters based on the context or input query, improving performance by adapting to the evolving needs of the task or user.
2. **[Contextual Compression](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/contextual_compression.ipynb)**: This technique reduces the size of the input context by distilling or summarizing key information, enabling more efficient retrieval and processing while retaining essential details for accurate response generation.
3. **[Document Augmentation](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/document_augmentation.py)**: This technique enriches retrieved documents by appending additional context or data from external sources, improving the quality and relevance of information fed to the model during generation.
4. **[Explainable Retrieval](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/explainable_retrieval.py)**: A retrieval technique focused on transparency, where the system provides clear explanations for why specific documents were retrieved, increasing trust and understanding in the retrieval process.
5. **[Fusion Retrieval](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/fusion_retrieval.py)**: A method that combines multiple retrieval techniques or models to aggregate and integrate information from diverse sources, resulting in a more comprehensive and accurate final response.
6. **[Graph RAG](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/graph_rag.py)**: This approach leverages graph-based structures to model relationships between retrieved documents or data points, allowing for more sophisticated reasoning and contextual understanding in the retrieval and generation process.
7. **[Hierarchical Indices](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/hierarchical_indices.py)**: A retrieval approach that organizes documents or data in a multi-level structure, allowing for more efficient and targeted searches by narrowing down relevant information through progressively refined layers of indexing.
8. **[Query Transformation](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/query_transformation.py)**: A technique that rewrites or reformulates the input query to improve retrieval performance, optimizing the search for relevant documents by making the query more effective for the underlying retrieval system.
9. **[Self-RAG](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/self_rag.ipynb)**: A method where the system autonomously retrieves and processes relevant documents during generation without external prompts, enabling more seamless integration of retrieval into the response creation process.
10. **[Semantic Chunking](https://github.com/sulaiman-shamasna/RAG-Techniques/blob/main/semantic_chunking.ipynb)**: A technique that breaks down large documents or data into smaller, semantically coherent chunks, allowing for more precise retrieval and enabling the model to focus on the most relevant sections during generation.

### References
1. [RAG Techniques](https://github.com/NirDiamant/RAG_Techniques/tree/main/all_rag_techniques)

RAG-Techniques

RAG-Techniques Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

RAG-Techniques

RAG-Techniques Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README