mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-04 13:20:12 +06:00

* feat: add colqwen2 (wip) * tests: fix test_attention_outputs * tests: reduce hidden size to accelerate tests * tests: fix `test_attention_outputs` 🥳 * fix: fix wrong parent class for `ColQwen2ForRetrievalOutput` * fix: minor typing and style changes * chore: run `make style` * feat: remove redundant `max_num_visual_tokens` attribute in `ColQwen2Processor` * tests: tweak comments * style: apply ruff formatter * feat: move default values for `visual_prompt_prefix` and `query_prefix` * docs: update ColQwen2 model card * docs: tweak model cards * docs: add required example config checkpoint * tests: update expected scores in integration test * docs: tweak quickstart snippets * fix: address PR comments * tests: fix colqwen2 tests + tweak comment in colpali test * tests: unskip useful tests * fix: fix bug when `visual_prompt_prefix` or `query_prefix` is an empty string * fix: fix ColPali outputs when `return_dict == False` * fix: fix issue with PaliGemma output not being a dict * docs: set default dtype to bfloat16 in quickstart snippets * fix: fix error when `return_dict=False` in ColPali and ColQwen2 * tests: fix special tokens not being replaced in input_ids * style: fix lint * fix: `ColQwen2Processor`'s `padding_side` is now set from `processor_config.json` * fix: remove unused `padding_side` in ColQwen2 model * docs: update ColQwen2's model doc * fix: fix harcoded vlm backbone class in ColQwen2Config * fix: remove `padding_side` from ColQwen2Processor as should fed from kwargs * docs: fix typo in model docstring * docs: add illuin mention in model docs * fix: let `padding_size` be handled by `tokenizer_config.json` * docs: add colpali reference url in colqwen2's model doc * docs: add Hf mention in model docs * docs: add late interaction mention in model docs * docs: tweak colqwen2 model doc * docs: update reference checkpoint for ColPali to v1.3 * docs: simplify quickstart snippets * docs: remove redundant `.eval()` * refactor: use `can_return_tuple` decorator for ColPali and ColQwen2 * docs: fix copyright date * docs: add missing copyright in tests * fix: raise error when `initializer_range` is not in config * docs: remove redundant `.eval()` in colpali doc * fix: fix `get_text_config` now that Qwen2VL has a proper `text_config` attribute See https://github.com/huggingface/transformers/pull/37268 for details about changes in Qwen2VL's config. * fix: add missing `initializer_range` attribute in `ColQwen2Config` * fix: use `get_text_config` in `resize_token_embeddings` * update colwen2 with auto_docstring * docs: fix wrong copyright year * chore: remove `raise` as `initializer_range` has a default value in `ColQwen2Config` * refactor: merge `inner_forward` into `forward` * Refactor colqwen2 after refactoring of qwen2VL, use modular for modeling code * protect torch import in modular to protect in processing * protect torch import in modular to protect in processing * tests: fix hf model path in ColQwen2 integration test * docs: clarify `attn_implementation` and add comments * docs: add fallback snippet for using offline PIL dummy images * docs: temporarily revert attn_implementation to `None` while sdpa is not fixed * docs: tweaks in colpali/colqwen2 quick start snippets * fix: add missing flags to enable SDPA/Flex Attention in ColQwen2 model * fix: add missing changes in modular file * fix modeling tests --------- Co-authored-by: yonigozlan <yoni.gozlan@huggingface.co>
264 lines
11 KiB
Python
264 lines
11 KiB
Python
# Copyright 2024 HuggingFace Inc.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
"""Testing suite for the ColPali processor."""
|
|
|
|
import shutil
|
|
import tempfile
|
|
import unittest
|
|
|
|
import torch
|
|
|
|
from transformers import GemmaTokenizer
|
|
from transformers.models.colpali.processing_colpali import ColPaliProcessor
|
|
from transformers.testing_utils import get_tests_dir, require_torch, require_vision
|
|
from transformers.utils import is_vision_available
|
|
|
|
from ...test_processing_common import ProcessorTesterMixin
|
|
|
|
|
|
if is_vision_available():
|
|
from transformers import (
|
|
ColPaliProcessor,
|
|
PaliGemmaProcessor,
|
|
SiglipImageProcessor,
|
|
)
|
|
|
|
SAMPLE_VOCAB = get_tests_dir("fixtures/test_sentencepiece.model")
|
|
|
|
|
|
@require_vision
|
|
class ColPaliProcessorTest(ProcessorTesterMixin, unittest.TestCase):
|
|
processor_class = ColPaliProcessor
|
|
|
|
@classmethod
|
|
def setUpClass(cls):
|
|
cls.tmpdirname = tempfile.mkdtemp()
|
|
image_processor = SiglipImageProcessor.from_pretrained("google/siglip-so400m-patch14-384")
|
|
image_processor.image_seq_length = 0
|
|
tokenizer = GemmaTokenizer(SAMPLE_VOCAB, keep_accents=True)
|
|
processor = PaliGemmaProcessor(image_processor=image_processor, tokenizer=tokenizer)
|
|
processor.save_pretrained(cls.tmpdirname)
|
|
|
|
@classmethod
|
|
def tearDownClass(cls):
|
|
shutil.rmtree(cls.tmpdirname, ignore_errors=True)
|
|
|
|
@require_torch
|
|
@require_vision
|
|
def test_process_images(self):
|
|
# Processor configuration
|
|
image_input = self.prepare_image_inputs()
|
|
image_processor = self.get_component("image_processor")
|
|
tokenizer = self.get_component("tokenizer", max_length=112, padding="max_length")
|
|
image_processor.image_seq_length = 14
|
|
|
|
# Get the processor
|
|
processor = self.processor_class(
|
|
tokenizer=tokenizer,
|
|
image_processor=image_processor,
|
|
)
|
|
|
|
# Process the image
|
|
batch_feature = processor.process_images(images=image_input, return_tensors="pt")
|
|
|
|
# Assertions
|
|
self.assertIn("pixel_values", batch_feature)
|
|
self.assertEqual(batch_feature["pixel_values"].shape, torch.Size([1, 3, 384, 384]))
|
|
|
|
@require_torch
|
|
@require_vision
|
|
def test_process_queries(self):
|
|
# Inputs
|
|
queries = [
|
|
"Is attention really all you need?",
|
|
"Are Benjamin, Antoine, Merve, and Jo best friends?",
|
|
]
|
|
|
|
# Processor configuration
|
|
image_processor = self.get_component("image_processor")
|
|
tokenizer = self.get_component("tokenizer", max_length=112, padding="max_length")
|
|
image_processor.image_seq_length = 14
|
|
|
|
# Get the processor
|
|
processor = self.processor_class(
|
|
tokenizer=tokenizer,
|
|
image_processor=image_processor,
|
|
)
|
|
|
|
# Process the image
|
|
batch_feature = processor.process_queries(text=queries, return_tensors="pt")
|
|
|
|
# Assertions
|
|
self.assertIn("input_ids", batch_feature)
|
|
self.assertIsInstance(batch_feature["input_ids"], torch.Tensor)
|
|
self.assertEqual(batch_feature["input_ids"].shape[0], len(queries))
|
|
|
|
# The following tests override the parent tests because ColPaliProcessor can only take one of images or text as input at a time.
|
|
|
|
def test_tokenizer_defaults_preserved_by_kwargs(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor_components["tokenizer"] = self.get_component("tokenizer", max_length=117, padding="max_length")
|
|
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
input_str = self.prepare_text_inputs()
|
|
inputs = processor(text=input_str, return_tensors="pt")
|
|
self.assertEqual(inputs[self.text_input_name].shape[-1], 117)
|
|
|
|
def test_image_processor_defaults_preserved_by_image_kwargs(self):
|
|
"""
|
|
We use do_rescale=True, rescale_factor=-1 to ensure that image_processor kwargs are preserved in the processor.
|
|
We then check that the mean of the pixel_values is less than or equal to 0 after processing.
|
|
Since the original pixel_values are in [0, 255], this is a good indicator that the rescale_factor is indeed applied.
|
|
"""
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor_components["image_processor"] = self.get_component(
|
|
"image_processor", do_rescale=True, rescale_factor=-1
|
|
)
|
|
processor_components["tokenizer"] = self.get_component("tokenizer", max_length=117, padding="max_length")
|
|
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
image_input = self.prepare_image_inputs()
|
|
|
|
inputs = processor(images=image_input, return_tensors="pt")
|
|
self.assertLessEqual(inputs[self.images_input_name][0][0].mean(), 0)
|
|
|
|
def test_kwargs_overrides_default_tokenizer_kwargs(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor_components["tokenizer"] = self.get_component("tokenizer", padding="longest")
|
|
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
input_str = self.prepare_text_inputs()
|
|
inputs = processor(text=input_str, return_tensors="pt", max_length=112, padding="max_length")
|
|
self.assertEqual(inputs[self.text_input_name].shape[-1], 112)
|
|
|
|
def test_kwargs_overrides_default_image_processor_kwargs(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor_components["image_processor"] = self.get_component(
|
|
"image_processor", do_rescale=True, rescale_factor=1
|
|
)
|
|
processor_components["tokenizer"] = self.get_component("tokenizer", max_length=117, padding="max_length")
|
|
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
image_input = self.prepare_image_inputs()
|
|
|
|
inputs = processor(images=image_input, do_rescale=True, rescale_factor=-1, return_tensors="pt")
|
|
self.assertLessEqual(inputs[self.images_input_name][0][0].mean(), 0)
|
|
|
|
def test_unstructured_kwargs(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
input_str = self.prepare_text_inputs()
|
|
inputs = processor(
|
|
text=input_str,
|
|
return_tensors="pt",
|
|
do_rescale=True,
|
|
rescale_factor=-1,
|
|
padding="max_length",
|
|
max_length=76,
|
|
)
|
|
|
|
self.assertEqual(inputs[self.text_input_name].shape[-1], 76)
|
|
|
|
def test_unstructured_kwargs_batched(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
image_input = self.prepare_image_inputs(batch_size=2)
|
|
inputs = processor(
|
|
images=image_input,
|
|
return_tensors="pt",
|
|
do_rescale=True,
|
|
rescale_factor=-1,
|
|
padding="longest",
|
|
max_length=76,
|
|
)
|
|
|
|
self.assertLessEqual(inputs[self.images_input_name][0][0].mean(), 0)
|
|
|
|
def test_doubly_passed_kwargs(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
image_input = self.prepare_image_inputs()
|
|
with self.assertRaises(ValueError):
|
|
_ = processor(
|
|
images=image_input,
|
|
images_kwargs={"do_rescale": True, "rescale_factor": -1},
|
|
do_rescale=True,
|
|
return_tensors="pt",
|
|
)
|
|
|
|
def test_structured_kwargs_nested(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
input_str = self.prepare_text_inputs()
|
|
|
|
# Define the kwargs for each modality
|
|
all_kwargs = {
|
|
"common_kwargs": {"return_tensors": "pt"},
|
|
"images_kwargs": {"do_rescale": True, "rescale_factor": -1},
|
|
"text_kwargs": {"padding": "max_length", "max_length": 76},
|
|
}
|
|
|
|
inputs = processor(text=input_str, **all_kwargs)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
|
|
self.assertEqual(inputs[self.text_input_name].shape[-1], 76)
|
|
|
|
def test_structured_kwargs_nested_from_dict(self):
|
|
if "image_processor" not in self.processor_class.attributes:
|
|
self.skipTest(f"image_processor attribute not present in {self.processor_class}")
|
|
processor_components = self.prepare_components()
|
|
processor = self.processor_class(**processor_components)
|
|
self.skip_processor_without_typed_kwargs(processor)
|
|
image_input = self.prepare_image_inputs()
|
|
|
|
# Define the kwargs for each modality
|
|
all_kwargs = {
|
|
"common_kwargs": {"return_tensors": "pt"},
|
|
"images_kwargs": {"do_rescale": True, "rescale_factor": -1},
|
|
"text_kwargs": {"padding": "max_length", "max_length": 76},
|
|
}
|
|
|
|
inputs = processor(images=image_input, **all_kwargs)
|
|
self.assertEqual(inputs[self.text_input_name].shape[-1], 76)
|