SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

fix typo (#9708)

Browse Source 
* fix typo

Co-authored-by: Suraj Patil <surajp815@gmail.com>
This commit is contained in:
Muennighoff
2021-01-21 09:21:01 +01:00
committed by GitHub
parent 4a20b7c450
commit 6a346f0358
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/transformers/modeling_tf_utils.py
View File

@@ -147,7 +147,7 @@ class TFCausalLanguageModelingLoss:
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy( loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(
from_logits=True, reduction=tf.keras.losses.Reduction.NONE from_logits=True, reduction=tf.keras.losses.Reduction.NONE
) )
# make sure only labels that are not equal to -100 do not affect loss # make sure only labels that are not equal to -100 affect the loss
active_loss = tf.not_equal(tf.reshape(labels, (-1,)), -100) active_loss = tf.not_equal(tf.reshape(labels, (-1,)), -100)
reduced_logits = tf.boolean_mask(tf.reshape(logits, (-1, shape_list(logits)[2])), active_loss) reduced_logits = tf.boolean_mask(tf.reshape(logits, (-1, shape_list(logits)[2])), active_loss)
labels = tf.boolean_mask(tf.reshape(labels, (-1,)), active_loss) labels = tf.boolean_mask(tf.reshape(labels, (-1,)), active_loss)