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

Fix processor kwargs qwen2 vl (#36890)

Browse Source 
* Fix qwen2_vl and qwen2_5_vl processors cutom images kwargs

* change version warning
This commit is contained in:
Yoni Gozlan
2025-03-24 13:19:26 -04:00
committed by GitHub
parent 48385aa4f4
commit 91455c1825
9 changed files with 66 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
tests/models/qwen2_5_vl/test_processor_qwen2_5_vl.py
View File

@@ -310,3 +310,19 @@ class Qwen2_5_VLProcessorTest(ProcessorTesterMixin, unittest.TestCase):
)
self.assertTrue(self.videos_input_name in out_dict_with_video)
self.assertEqual(len(out_dict_with_video[self.videos_input_name]), 71280)
def test_kwargs_overrides_custom_image_processor_kwargs(self):
processor_components = self.prepare_components()
processor_components["image_processor"] = self.get_component("image_processor")
processor_components["tokenizer"] = self.get_component("tokenizer")
processor_kwargs = self.prepare_processor_dict()
processor = self.processor_class(**processor_components, **processor_kwargs, use_fast=True)
self.skip_processor_without_typed_kwargs(processor)
input_str = self.prepare_text_inputs()
image_input = self.prepare_image_inputs()
inputs = processor(text=input_str, images=image_input, max_pixels=56 * 56 * 4, return_tensors="pt")
self.assertEqual(inputs[self.images_input_name].shape[0], 612)
inputs = processor(text=input_str, images=image_input, return_tensors="pt")
self.assertEqual(inputs[self.images_input_name].shape[0], 800)

16
tests/models/qwen2_vl/test_processor_qwen2_vl.py
View File

@@ -307,3 +307,19 @@ class Qwen2VLProcessorTest(ProcessorTesterMixin, unittest.TestCase):
)
self.assertTrue(self.videos_input_name in out_dict_with_video)
self.assertEqual(len(out_dict_with_video[self.videos_input_name]), 71280)
def test_kwargs_overrides_custom_image_processor_kwargs(self):
processor_components = self.prepare_components()
processor_components["image_processor"] = self.get_component("image_processor")
processor_components["tokenizer"] = self.get_component("tokenizer")
processor_kwargs = self.prepare_processor_dict()
processor = self.processor_class(**processor_components, **processor_kwargs, use_fast=True)
self.skip_processor_without_typed_kwargs(processor)
input_str = self.prepare_text_inputs()
image_input = self.prepare_image_inputs()
inputs = processor(text=input_str, images=image_input, return_tensors="pt")
self.assertEqual(inputs[self.images_input_name].shape[0], 800)
inputs = processor(text=input_str, images=image_input, max_pixels=56 * 56 * 4, return_tensors="pt")
self.assertEqual(inputs[self.images_input_name].shape[0], 612)