scoriiu

5 ms·

1. 1.▲
   by scoriiu 2mo ago

   that's a clean answer thanks. the 85% rule for BKT routing is more principled than most. the one thing i'd still watch is wether summit mastery still holds or decays over time, since BKT can't tell "build the schema&quot
2. 2.▲
   by scoriiu 2mo ago

   curious how the tutor handles a kid who's confidently wrong. from the description the ladder kicks in at the first sign of struggle, which sounds like it's built to prevent mistakes rather than use them. kapur's productive fa
3. 3.▲
   by scoriiu 2mo ago

   did you skip simd just because the model's tiny? naive conv perf is honestly the only reason i haven't done exactly this for the cnn
4. 4.▲
   by scoriiu 2mo ago

   Should be very doable. I ship a small CNN in a browser extension via onnxruntime-web and the model weights were never the bottleneck, the runtime was. The wasm backend adds a few MB of runtime before your first inference, so a 500kb model w
5. 5.▲
   A free chess trainer where the coach explains your games in plain English(coachess.app)

   2 points by scoriiu 3mo ago|0 comments