explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2024-12-25 17:36:30 +03:00
Code Issues Packages Projects Releases Wiki Activity

fix comments

Browse Source
This commit is contained in:
svlandeg 2020-06-02 19:50:44 +02:00
parent 6208d322d3
commit ef834b4cd7
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
spacy/util.py
View File

@ -682,23 +682,23 @@ def minibatch_by_words(examples, size, count_words=len, tolerance=0.2, discard_o
if not discard_oversize:
yield [example]
# add the example to the current batch if it still fits and there's no overflow yet
# add the example to the current batch if there's no overflow yet and it still fits
elif overflow_size == 0 and (current_size + n_words) < target_size:
batch.append(example)
current_size += n_words
# add the example to the overflow buffer if it fits in the tolerance margins
# add the example to the overflow buffer if it fits in the tolerance margin
elif (current_size + overflow_size + n_words) < (target_size + tol_size):
overflow.append(example)
overflow_size += n_words
# yield the previous batch and start a new one
# yield the previous batch and start a new one. The new one gets the overflow examples.
else:
yield batch
target_size = next(size_)
tol_size = target_size * tolerance
# In theory it may happen that the current example now exceeds the new target_size,
# but that seems like an unimportant edge case if batch sizes are variable anyway?
# In theory it may happen that the current example + overflow examples now exceed the new
# target_size, but that seems like an unimportant edge case if batch sizes are variable?
batch = overflow
batch.append(example)
current_size = overflow_size + n_words