open-spaced-repetition/fsrs-rs

Issues

• [BUG] `memory_state_from_sm2` still uses the forgetting curve funtion of FSRSv4.

  #190 opened 7 months ago by L-M-Sherlock
  0

• Use median in calculating cost

  #188 opened 8 months ago by user1823
  1

• [BUG] Training is 2.5x slower in burn-0.13.0 than burn-0.12.1

  #185 opened 8 months ago by L-M-Sherlock
  3

• Tests should be in a tests folder

  #178 opened 9 months ago by Kees-van-Beilen
  2

• Add a example file

  #144 opened a year ago by SwayStar123
  4

• Better outlier filter for trainset

  #121 opened 10 months ago by user1823
  25

• support WASM

  #99 opened a year ago by hxse
  4

• Pre-training Only when the number of reviews is less than 1000

  #152 opened a year ago by L-M-Sherlock
  5

• Reference usage?

  #151 opened a year ago by aehlke
  5

• User guide

  #133 opened a year ago by jawadkho
  3

• Inference.rs uses the new power curve, but the default parameters are from v4

  #143 opened a year ago by Expertium
  17

• Add an option to turn off outlier filter when benchmark

  #136 opened a year ago by L-M-Sherlock
  1

• What's the difference between this repo and rs-fsrs?

  #130 opened a year ago by Expertium
  1

• Skip reviews with time = 0 when calculating average answer times

  #126 opened a year ago by user1823
  1

• Enhancement: Include incomplete revlogs even when training

  #114 opened a year ago by user1823
  4

• TODO: speed up finding optimal retention via Brent's method

  #118 opened a year ago by L-M-Sherlock
  0

• Achieve parity with the Python optimizer

  #88 opened a year ago by dae
  10

• Support for Online Learning?

  #56 opened a year ago by user1823
  3

• Consider time-frame limitation?

  #117 opened a year ago by dae
  3

• [Question] How to choose "Days to simulate"?

  #81 opened a year ago by Expertium
  14

• Request: Ignore reviews before "Forget"

  #113 opened a year ago by user1823
  9

• [Enhancement] Use more splits while training with larger datasets

  #106 opened a year ago by user1823
  3

• [BUG] Huge difference between Rust optimizer and Python optimizer

  #78 opened a year ago by L-M-Sherlock
  15

• [Feature Request[ Use two different sets of initial parameters, then average out the results

  #84 opened a year ago by Expertium
  4

• Use the first revlog in the "known" review history for converting SM-2 ivl & ease to memory states

  #87 opened a year ago by user1823
  13

• [BUG] Potential inconsistency in optimal_retention.rs

  #79 opened a year ago by Expertium
  20

• Review logs that requires special treatment when calculating the memory states from review history

  #63 opened a year ago by L-M-Sherlock
  37

• [TODO] feature: extract optimal retention parameters from revlog

  #72 opened a year ago by L-M-Sherlock
  0

• warning in compile

  #41 opened a year ago by asukaminato0721
  3

• [TODO] feature: use real Answer Buttons probability

  #61 opened a year ago by L-M-Sherlock
  1

• [TODO] feature: support scheduler

  #57 opened a year ago by L-M-Sherlock
  0

• -burn -> -rs

  #54 opened a year ago by dae
  2

• Integrate into Anki

  #27 opened a year ago by dae
  9

• [TODO] feature: evaluate the model

  #51 opened a year ago by L-M-Sherlock
  0

• [TODO] feature: search optimal retention

  #45 opened a year ago by L-M-Sherlock
  0

• Add license

  #43 opened a year ago by dae
  10

• [TODO] feature: pre-training for initial stability

  #40 opened a year ago by L-M-Sherlock
  0

• Remove reshape() calls

  #28 opened a year ago by dae
  1

• Question about RevlogEntry mapping

  #37 opened a year ago by dae
  5

• [TODO] feature: freeze weights

  #20 opened a year ago by L-M-Sherlock
  10

• Learning Rust

  #6 opened a year ago by cjdduarte
  3

• [TODO] feature: weight clipper

  #4 opened a year ago by L-M-Sherlock
  6

• Questions about FSRSDataset

  #34 opened a year ago by dae
  6

• [TODO] feature: build dataset from Anki's sqlite database

  #3 opened a year ago by L-M-Sherlock
  17

• [TODO] feature: custom sampler for batch

  #22 opened a year ago by L-M-Sherlock
  9

• [TODO] feature: CosineAnnealingLR

  #21 opened a year ago by L-M-Sherlock
  3

• [TODO] fix: ndarray broadcast error in mini-batch

  #24 opened a year ago by L-M-Sherlock
  0

• [TODO] feature: mini-batch & sequence-padding

  #5 opened a year ago by L-M-Sherlock
  1

• [TODO] fix: BCELoss

  #17 opened a year ago by L-M-Sherlock
  0

• [TODO] rewrite FSRSItem

  #7 opened a year ago by L-M-Sherlock
  0