Skip to content
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

SCCE fix for bug in Jax<0.2.7 #130

Merged
merged 3 commits into from
Dec 22, 2020
Merged

SCCE fix for bug in Jax<0.2.7 #130

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
03443f5
fix for a bug in jax<0.2.7
alexander-g Dec 18, 2020
eb7b1dc
test case +black
alexander-g Dec 19, 2020
29b3d82
jax version update
alexander-g Dec 21, 2020
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
14 changes: 13 additions & 1 deletion elegy/losses/sparse_categorical_crossentropy_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,9 @@
import elegy


import jax.numpy as jnp
import jax, jax.numpy as jnp
import numpy as np
import tensorflow.keras as keras


#
Expand Down Expand Up @@ -47,3 +49,13 @@ def test_scce_out_of_bounds():
scce = elegy.losses.SparseCategoricalCrossentropy(check_bounds=False)
assert not jnp.isnan(scce(ytrue0, ypred)).any()
assert not jnp.isnan(scce(ytrue1, ypred)).any()


def test_scce_uint8_ytrue():
ypred = np.random.random([2, 256, 256, 10])
ytrue = np.random.randint(0, 10, size=(2, 256, 256)).astype(np.uint8)

loss0 = elegy.losses.sparse_categorical_crossentropy(ytrue, ypred, from_logits=True)
loss1 = keras.losses.sparse_categorical_crossentropy(ytrue, ypred, from_logits=True)

assert np.allclose(loss0, loss1)
46 changes: 37 additions & 9 deletions poetry.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.