Processing a whole book downloads all pages and runs the OCR in parallel processes. This is like a Denial-of-Service attack for the image server and creates also a huge load on the OCR server with potential out-of-memory errors.
Parallel processing is fine, but should work with a predefined number of worker processes.
csidirop
commented
1 year ago
Processing a whole book downloads all pages and runs the OCR in parallel processes. This is like a Denial-of-Service attack for the image server and creates also a huge load on the OCR server with potential out-of-memory errors.
Parallel processing is fine, but should work with a predefined number of worker processes.