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

⚠️ Raise Exception when trying to generate 0 tokens ⚠️ (#28621)

Browse Source 
* change warning to exception

* Update src/transformers/generation/utils.py

Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* validate `max_new_tokens` > 0 in `GenerationConfig`

* fix truncation test parameterization in `TextGenerationPipelineTests`

---------

Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>
This commit is contained in:
Daniel Korat
2024-02-07 14:42:01 +02:00
committed by GitHub
parent 349a6e8542
commit abf8f54a01
3 changed files with 9 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/pipelines/test_pipelines_text_generation.py
View File

@@ -93,17 +93,19 @@ class TextGenerationPipelineTests(unittest.TestCase):
## -- test tokenizer_kwargs
test_str = "testing tokenizer kwargs. using truncation must result in a different generation."
input_len = len(text_generator.tokenizer(test_str)["input_ids"])
output_str, output_str_with_truncation = (
text_generator(test_str, do_sample=False, return_full_text=False)[0]["generated_text"],
text_generator(test_str, do_sample=False, return_full_text=False, min_new_tokens=1)[0]["generated_text"],
text_generator(
test_str,
do_sample=False,
return_full_text=False,
min_new_tokens=1,
truncation=True,
max_length=3,
max_length=input_len + 1,
)[0]["generated_text"],
)
assert output_str != output_str_with_truncation # results must be different because one hd truncation
assert output_str != output_str_with_truncation # results must be different because one had truncation
# -- what is the point of this test? padding is hardcoded False in the pipeline anyway
text_generator.tokenizer.pad_token_id = text_generator.model.config.eos_token_id