sort FSRSItem by length to speed up training

open-spaced-repetition · Aug 25, 2023 · e25a7a8 · e25a7a8
1 parent 7d80f4e
commit e25a7a8
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 4 deletions.
diff --git a/src/convertor.rs b/src/convertor.rs
@@ -1,5 +1,6 @@
 use chrono::prelude::*;
 use chrono_tz::Tz;
+use itertools::Itertools;
 use rusqlite::{Connection, Result, Row};
 use std::collections::HashMap;
 
@@ -192,11 +193,13 @@ fn convert_to_fsrs_items(
 pub fn anki_to_fsrs() -> Vec<FSRSItem> {
  let revlogs = read_collection().expect("read error");
  let revlogs_per_card = group_by_cid(revlogs);
- revlogs_per_card
+ let mut revlogs = revlogs_per_card
  .into_iter()
  .filter_map(|entries| convert_to_fsrs_items(entries, 4, Tz::Asia__Shanghai))
  .flatten()
- .collect()
+ .collect_vec();
+ revlogs.sort_by_key(|r| r.reviews.len());
+ revlogs
 }
 
 #[cfg(test)]

diff --git a/src/training.rs b/src/training.rs
@@ -117,13 +117,13 @@ pub fn train<B: ADBackend<FloatElem = f32>>(
 
  let dataloader_train = DataLoaderBuilder::new(batcher_train)
  .batch_size(config.batch_size)
- .shuffle(config.seed)
+ // .shuffle(config.seed)
  .num_workers(config.num_workers)
  .build(FSRSDataset::train());
 
  let dataloader_test = DataLoaderBuilder::new(batcher_valid)
  .batch_size(config.batch_size)
- .shuffle(config.seed)
+ // .shuffle(config.seed)
  .num_workers(config.num_workers)
  .build(FSRSDataset::test());