CAG

Cache-Augmented Generation: A Simple, Efficient Alternative to RAG

1030

149

Python

Cache-Augmented Generation (CAG)

Retrieval-Augmented Generation (RAG) has emerged as a powerful approach for enhancing language models by integrating external knowledge sources. However, RAG also introduces several challenges, including:

Retrieval Latency – Delays caused by real-time retrieval steps.
Retrieval Errors – Inaccuracies in selecting relevant documents.
System Complexity – Increased architectural and maintenance overhead.

To address these limitations, we propose Cache-Augmented Generation (CAG)—an alternative paradigm that bypasses real-time retrieval. CAG leverages the extended context windows of modern large language models (LLMs) by preloading all relevant resources into the model’s context and caching its runtime parameters. During inference, the preloaded KV-cache enables the model to generate responses directly, eliminating the need for retrieval.

Advantages of CAG

Reduced Latency – Eliminates real-time retrieval, enabling faster inference.
Improved Reliability – Minimizes retrieval errors while maintaining context relevance.
Simplified Design – Provides a streamlined, retrieval-free alternative to RAG, achieving comparable or superior results with lower complexity.

Limitations of CAG

Limited Knowledge Size – CAG requires the entire knowledge source to fit within the context window, making it less suitable for tasks involving extremely large datasets.
Context Length Constraints – The performance of LLMs may degrade with very long contexts (reference).

Our paper, which will be presented at the ACM Web Conference 2025 as a short paper, investigates the relationship between model performance and context length, providing insights into scenarios where CAG excels.

The limitations of CAG are rapidly being addressed by advancements in LLMs with longer context windows and improved capabilities for extracting relevant information from extended inputs. As these models continue to evolve, CAG is expected to handle increasingly complex applications, making it a practical and scalable alternative to traditional RAG.

Installation

pip install -r ./requirements.txt

Preparation

[!IMPORTANT]
download the required squad and hotpotqa datasets by curl script
sh ./downloads.sh

[!IMPORTANT]
create .env file by .env.template and input the keys required
cp ./.env.template ./.env

Usage

rag.py is for RAG Experiment
kvcache.py is for CAG Experiment

Parameter Usage – kvcache.py

--kvcache: “file”
--dataset: “hotpotqa-train” or “squad-train”
--similarity “bertscore”
--modelname: “meta-llama/Llama-3.1-8B-Instruct”
--maxKnowledge: “”, int, select how many document in dataset, explanation in Note
--maxParagraph: 100
--maxQuestion int, max question number, explanation in Note
--randomSeed: “”, int, a random seed number
--output: “”, str, output filepath string
--usePrompt, add this parameter if not using CAG knowledge cache acceleration

Example – kvcache.py

python ./kvcache.py --kvcache file --dataset "squad-train" --similarity bertscore \
    --maxKnowledge 5 --maxParagraph 100 --maxQuestion 1000  \
    --modelname "meta-llama/Llama-3.1-8B-Instruct" --randomSeed 0 \
    --output "./result_kvcache.txt"

Parameter Usage – rag.py

--index: “openai” or “bm25”
--dataset: “hotpotqa-train” or “squad-train”
--similarity “bertscore”
--maxKnowledge: “”, int, select how many document in dataset, explanation in Note
--maxParagraph: 100
--maxQuestion int, max question number, explanation in Note
--topk: int, the similarity topk of retrieval
--modelname: “meta-llama/Llama-3.1-8B-Instruct”
--randomSeed: “”, int, a random seed number
--output: “”, str, output filepath string

Example – rag.py

python ./rag.py --index "bm25" --dataset "hotpotqa-train" --similarity bertscore \
    --maxKnowledge 80 --maxParagraph 100 --maxQuestion 80 --topk 3 \
    --modelname "meta-llama/Llama-3.1-8B-Instruct" --randomSeed  0 \
    --output  "./rag_results.txt"

Note:

`--maxKnowledge` parameter notice:

[!NOTE]
Approximate Tokens count corresponding to knowledge document size of “squad-train” and “hotpotqa-train” dataset.

datasets=(“squad-train”)

when k = 3, tokens = 21,000

when k = 4, tokens = 32,000

when k = 7, tokens = 50,000

datasets=(“hotpotqa-train”)

all k = 7405 article, tokens = 10,038,084

when k = 1, tokens = 1,400

when k = 16, tokens = 22,400

when k = 24, tokens = 33,667

when k = 32, tokens = 44,800

when k = 48, tokens = 64,000

when k = 64, tokens = 85,000

when k = 80, tokens = 106,000

`--maxQuestion` parameter notice:

when using “squad-train” dataset, 1 knowledge has average 150 questions

when using “hotpotqa-train” dataset, 1 knowledge has 1 question

[!TIP]
Since 1 document in “hotpoqa-train” dataset has only 1 question, it may not satisfy large-scale evaluation.
Multiple evaluation could be a relatively better approach.

`Docker`

To build the docker image, run

 docker build -t my-cag-app .

and to run the container, run this for GPU users

docker run --gpus all -it --rm my-cag-app

docker run -it --rm my-cag-app

for CPU users.

if the .env file details were empty while building you will get error similar to this below

Traceback (most recent call last):
  File "/app/./kvcache.py", line 35, in <module>
    env = validate_env_variables()
          ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/app/./kvcache.py", line 31, in validate_env_variables
    raise ValueError(f"Missing required environment variable: {key}")
ValueError: Missing required environment variable: HF_TOKEN

so ensure you populate the .env file before building the docker image

Note that the he CMD directive in the Dockerfile runs the kvcache.py script by default. You can override this in the docker run command if you’d like to execute other scripts like rag.py. For example:

docker run --gpus all -it --rm my-cag-app python ./rag.py --index "bm25" --dataset "hotpotqa-train" --similarity bertscore --maxKnowledge 80 --maxParagraph 100 --maxQuestion 80 --topk 3 --modelname "meta-llama/Llama-3.1-8B-Instruct" --randomSeed 0 --output "./rag_results.txt"

Citation

@misc{chan2024dontragcacheaugmentedgeneration,
      title={Don't Do RAG: When Cache-Augmented Generation is All You Need for Knowledge Tasks}, 
      author={Brian J Chan and Chao-Ting Chen and Jui-Hung Cheng and Hen-Hsen Huang},
      year={2024},
      eprint={2412.15605},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2412.15605}, 
}

Acknowledgments

This work was partially supported by National Science and Technology Council (NSTC), Taiwan, under the grant 112-2221-E-001-016-MY3, by Academia Sinica, under the grant 236d-1120205, and by National Center for High-performance Computing (NCHC), National Applied Research Laboratories (NARLabs), and NSTC under the project “Taiwan’s 113th year endeavoring in the promotion of a trustworthy generative AI large language model and the cultivation of literacy capabilities (Trustworthy AI Dialog Engine, TAIDE)”.
We sincerely thank Discover AI and the many individuals who have introduced, shared, and discussed our work, contributing to its broader visibility and impact.

CAG