davidzollo commented on PR #10578: URL: https://github.com/apache/seatunnel/pull/10578#issuecomment-4915063857
Thanks @corgy-w for working on batched LLM and Embedding transforms. This is a useful direction for reducing per-record model-call overhead and making these transforms more practical. After rechecking the current `dev` branch, batching support is already present through `process_batch_size` / `inference_batch_size` in the NLP model transform configuration. This draft has also become stale enough that continuing it as-is would be hard to review safely, so I'm closing this PR now. If checkpoint-state behavior is still missing or incomplete on the latest `dev` branch, a smaller fresh PR focused only on that remaining gap would be very welcome. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
