ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update tokenizer_utils.py

Browse source
This commit is contained in:
Daniel Han-Chen 2024-03-24 19:32:54 +11:00
commit 7e1f9118e7
1 changed files with 3 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
unsloth/tokenizer_utils.py
View file

@ -76,7 +76,6 @@ pass
def get_sorted_dict(dictionary):
# Get sorted dict by values 0, 1, 2, ...
sorted_keys = sorted(dictionary.values())
inverted_dictionary = { value : key for key, value in dictionary.items() }
@ -164,7 +163,9 @@ def convert_to_fast_tokenizer(
# Now load it!
fast_tokenizer = AutoTokenizer.from_pretrained(new_location)
return fast_tokenizer
if assert_same_tokenization(slow_tokenizer, fast_tokenizer):
return fast_tokenizer
return slow_tokenizer
pass