Skip to content

Conversation

@ssubhanjali
Copy link

  • Added process_qa_chunk() function to handle parallel processing on individual GPUs
  • Modified make_dataset() to distribute QA pairs across available GPUs using multiprocessing
  • Each GPU process creates its own embedding model and backend directory
  • Added per-GPU progress bars for real-time monitoring

@puririshi98
Copy link
Contributor

please fix precommit and ensure that this incurs no acc change between master and this branch when tested on ralphs data

@puririshi98
Copy link
Contributor

please also update ChangeLog.MD to mention that you have accelerated the retrieval in examples/llm/txt2kg_rag.py using multiproc

@ssubhanjali ssubhanjali force-pushed the ss/multigpu branch 5 times, most recently from d8a9435 to 43ca4f0 Compare October 28, 2025 07:07
* Added process_qa_chunk() function to handle parallel processing on individual GPUs
* Modified make_dataset() to distribute QA pairs across available GPUs using multiprocessing
* Each GPU process creates its own embedding model and backend directory
* Added per-GPU progress bars for real-time monitoring
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants