Skip to content

Relax MNIST convergence canary threshold (97% → 96%) #26

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ausimian merged 1 commit into from
Apr 16, 2026
Merged

Relax MNIST convergence canary threshold (97% → 96%) #26

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions RELEASE.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,11 @@
# Release notes for next release

## Fixed

- Relax MNIST convergence canary threshold from 97% to 96% to eliminate
stochastic flaps (observed 96.99% on occasional runs). The test is a
sanity gate, not a performance benchmark.

## Added

- M14 — Serving concurrency: stream-per-process. `Emily.Stream` lets
Expand Down
4 changes: 2 additions & 2 deletions test/emily/training/mnist_full_test.exs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ defmodule Emily.Training.MnistFullTest do

Opt-in — `mix test --only training_full`. Downloads MNIST via
`scidata` (~11 MB, cached in `~/Library/Caches/scidata`), trains an
Axon MLP on `Emily.Compiler`, and asserts >97% test accuracy.
Axon MLP on `Emily.Compiler`, and asserts >96% test accuracy.

Purpose: catch systemic numerical drift that the curve-matching
tests in `mlp_curve_test.exs` can't see — those compare Emily vs
Expand Down Expand Up @@ -33,7 +33,7 @@ defmodule Emily.Training.MnistFullTest do

@batch_size 128
@epochs 5
@target_accuracy 0.97
@target_accuracy 0.96

test "Axon MLP reaches >#{trunc(@target_accuracy * 100)}% test accuracy under Emily.Compiler" do
{train_batches, test_images, test_labels} = load_mnist(@batch_size)
Expand Down
Loading