Open connortbot opened 5 days ago
Current back lambda, even at 900 seconds doesn't go generate the stories fast enough even with a low bar like 7 subjects, 6 CEFR levels, 1 language.
Potential solution is to multithread the generation on the lambda, given that the timing is bottlenecked by API calls and not performance.
description
Current back lambda, even at 900 seconds doesn't go generate the stories fast enough even with a low bar like 7 subjects, 6 CEFR levels, 1 language.
details
Potential solution is to multithread the generation on the lambda, given that the timing is bottlenecked by API calls and not performance.