Closed joshjchayes closed 1 year ago
We should be able to run batches in parallel, since they are independent calculations. This should speed up larger retrievals significantly.
Pretty much all this will go in Retriever.run_batched_retrieval()
Retriever.run_batched_retrieval()
I think this is doable in the same way I implemented running your code in parallel with the partial function and pool.
We should be able to run batches in parallel, since they are independent calculations. This should speed up larger retrievals significantly.
Pretty much all this will go in
Retriever.run_batched_retrieval()