Add DeepSeek integration blog #3595
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Fanit Kolchina <[email protected]>
kolchfa-aws
requested review from
elfisher,
AMoo-Miki,
nknize,
krisfreedain,
peterzhuamazon,
CEHENKLE,
dtaivpp,
nateynateynate and
natebower
as code owners
Signed-off-by: Fanit Kolchina <[email protected]>
natebower
reviewed
Jan 30, 2025
There was a problem hiding this comment.
@kolchfa-aws Editorial review complete. Please see my comments and changes and let me know if you have any questions. Thanks!
| redirect_from: '/authors/vikash/' | ||
| --- | ||
|
|
||
| **Vikash Tiwari** is a Senior Software Development Engineer at Amazon Web Services, specializing in OpenSearch Vector Search. He is passionate about distributed systems, large-scale machine learning, scalable search architectures, and database internals. His expertise spans vector search, indexing optimizations, and efficient data retrieval, and he is deeply interested in learning and enhancing modern database systems. |
There was a problem hiding this comment.
Suggested change
| **Vikash Tiwari** is a Senior Software Development Engineer at Amazon Web Services, specializing in OpenSearch Vector Search. He is passionate about distributed systems, large-scale machine learning, scalable search architectures, and database internals. His expertise spans vector search, indexing optimizations, and efficient data retrieval, and he is deeply interested in learning and enhancing modern database systems. | |
| **Vikash Tiwari** is a Senior Software Development Engineer at AWS specializing in OpenSearch vector search. He is passionate about distributed systems, large-scale machine learning, scalable search architectures, and database internals. His expertise spans vector search, indexing optimizations, and efficient data retrieval, and he is deeply interested in learning and enhancing modern database systems. |
|
|
||
| ## RAG with OpenSearch and DeepSeek | ||
|
|
||
| The following diagram presents the RAG workflow in OpenSearch using the DeepSeek model. |
There was a problem hiding this comment.
Suggested change
| The following diagram presents the RAG workflow in OpenSearch using the DeepSeek model. | |
| The following diagram depicts the RAG workflow in OpenSearch using the DeepSeek model. |
| pip install opensearch-py transformers torch sentence-transformers | ||
| ``` | ||
|
|
||
| These packages form the backbone of our RAG system: |
There was a problem hiding this comment.
Suggested change
| These packages form the backbone of our RAG system: | |
| These packages form the backbone of the RAG system: |
|
|
||
| These packages form the backbone of our RAG system: | ||
|
|
||
| * `opensearch-py`: The official Python client for OpenSearch. |
There was a problem hiding this comment.
Suggested change
| * `opensearch-py`: The official Python client for OpenSearch. | |
| * `opensearch-py`: The official Python client for OpenSearch |
| These packages form the backbone of our RAG system: | ||
|
|
||
| * `opensearch-py`: The official Python client for OpenSearch. | ||
| * `transformers`: Hugging Face's library for working with transformer models. |
There was a problem hiding this comment.
Suggested change
| * `transformers`: Hugging Face's library for working with transformer models. | |
| * `transformers`: Hugging Face's library for working with transformer models |
|
|
||
| In this step, you'll initialize two key components: | ||
|
|
||
| - The SentenceTransformer model for creating document embeddings |
There was a problem hiding this comment.
Suggested change
| - The SentenceTransformer model for creating document embeddings | |
| - The SentenceTransformer model for creating document embeddings. |
| In this step, you'll initialize two key components: | ||
|
|
||
| - The SentenceTransformer model for creating document embeddings | ||
| - The DeepSeek model for generating responses. The MiniLM-L6-v2 model provides a good balance between performance and accuracy for embeddings. |
There was a problem hiding this comment.
Suggested change
| - The DeepSeek model for generating responses. The MiniLM-L6-v2 model provides a good balance between performance and accuracy for embeddings. | |
| - The DeepSeek model for generating responses. The MiniLM-L6-v2 model provides a good balance between embedding performance and accuracy. |
| The RAG pipeline consists of two main functions: | ||
|
|
||
| * `index_document`: Converts text into embeddings and stores them in OpenSearch. | ||
| * `query_documents`: Performs similarity search using k-NN to find relevant documents. In this example, k-NN search returns the 3 most similar documents; you can adjust this number based on your needs. |
There was a problem hiding this comment.
Suggested change
| * `query_documents`: Performs similarity search using k-NN to find relevant documents. In this example, k-NN search returns the 3 most similar documents; you can adjust this number based on your needs. | |
| * `query_documents`: Performs similarity search using k-NN to find relevant documents. In this example, k-NN search returns the three most similar documents; you can adjust this number based on your needs. |
|
|
||
| ## Try it out | ||
|
|
||
| The following complete one-click script contains all preceding steps. Save this script as `deepseek_rag.py`: |
There was a problem hiding this comment.
Suggested change
| The following complete one-click script contains all preceding steps. Save this script as `deepseek_rag.py`: | |
| The following complete one-click script contains all of the preceding steps. Save this script as `deepseek_rag.py`: |
|
|
||
| This simple RAG implementation combines the power of OpenSearch's vector search capabilities with DeepSeek's advanced language understanding. While this is a basic setup, it provides a foundation that you can build upon for more complex applications. | ||
|
|
||
| Key benefits of this implementation: |
There was a problem hiding this comment.
Suggested change
| Key benefits of this implementation: | |
| The following are some key benefits of this implementation: |
|
@nateynateynate This should not be merged until @kolchfa-aws addresses editorial review. Thanks! |
|
Shit. Sorry, I'll revert. |
|
@kolchfa-aws - My apologies. You might have to file another PR to get this completed. I can't seem to find a button to reopen it. |
1 task
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Closes #3587
Check List
By submitting this pull request, I confirm that my contribution is made under the terms of the BSD-3-Clause License.