Spaces:
Paused
Paused
File size: 6,609 Bytes
ee6e328 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 |
# Copyright 2022 The HuggingFace Team. All rights reserved.
#
# 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.
import unittest
from transformers import MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING, is_vision_available
from transformers.pipelines import pipeline
from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_gpu,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_torch_available():
import torch
if is_vision_available():
from PIL import Image
else:
class Image:
@staticmethod
def open(*args, **kwargs):
pass
@is_pipeline_test
@require_torch
@require_vision
class VisualQuestionAnsweringPipelineTests(unittest.TestCase):
model_mapping = MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING
def get_test_pipeline(self, model, tokenizer, processor):
vqa_pipeline = pipeline("visual-question-answering", model="hf-internal-testing/tiny-vilt-random-vqa")
examples = [
{
"image": Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png"),
"question": "How many cats are there?",
},
{
"image": "./tests/fixtures/tests_samples/COCO/000000039769.png",
"question": "How many cats are there?",
},
]
return vqa_pipeline, examples
def run_pipeline_test(self, vqa_pipeline, examples):
outputs = vqa_pipeline(examples, top_k=1)
self.assertEqual(
outputs,
[
[{"score": ANY(float), "answer": ANY(str)}],
[{"score": ANY(float), "answer": ANY(str)}],
],
)
@require_torch
def test_small_model_pt(self):
vqa_pipeline = pipeline("visual-question-answering", model="hf-internal-testing/tiny-vilt-random-vqa")
image = "./tests/fixtures/tests_samples/COCO/000000039769.png"
question = "How many cats are there?"
outputs = vqa_pipeline(image=image, question="How many cats are there?", top_k=2)
self.assertEqual(
outputs, [{"score": ANY(float), "answer": ANY(str)}, {"score": ANY(float), "answer": ANY(str)}]
)
outputs = vqa_pipeline({"image": image, "question": question}, top_k=2)
self.assertEqual(
outputs, [{"score": ANY(float), "answer": ANY(str)}, {"score": ANY(float), "answer": ANY(str)}]
)
@require_torch
@require_torch_gpu
def test_small_model_pt_blip2(self):
vqa_pipeline = pipeline(
"visual-question-answering", model="hf-internal-testing/tiny-random-Blip2ForConditionalGeneration"
)
image = "./tests/fixtures/tests_samples/COCO/000000039769.png"
question = "How many cats are there?"
outputs = vqa_pipeline(image=image, question=question)
self.assertEqual(outputs, [{"answer": ANY(str)}])
outputs = vqa_pipeline({"image": image, "question": question})
self.assertEqual(outputs, [{"answer": ANY(str)}])
outputs = vqa_pipeline([{"image": image, "question": question}, {"image": image, "question": question}])
self.assertEqual(outputs, [[{"answer": ANY(str)}]] * 2)
vqa_pipeline = pipeline(
"visual-question-answering",
model="hf-internal-testing/tiny-random-Blip2ForConditionalGeneration",
model_kwargs={"torch_dtype": torch.float16},
device=0,
)
self.assertEqual(vqa_pipeline.model.device, torch.device(0))
self.assertEqual(vqa_pipeline.model.language_model.dtype, torch.float16)
self.assertEqual(vqa_pipeline.model.vision_model.dtype, torch.float16)
outputs = vqa_pipeline(image=image, question=question)
self.assertEqual(outputs, [{"answer": ANY(str)}])
@slow
@require_torch
def test_large_model_pt(self):
vqa_pipeline = pipeline("visual-question-answering", model="dandelin/vilt-b32-finetuned-vqa")
image = "./tests/fixtures/tests_samples/COCO/000000039769.png"
question = "How many cats are there?"
outputs = vqa_pipeline(image=image, question=question, top_k=2)
self.assertEqual(
nested_simplify(outputs, decimals=4), [{"score": 0.8799, "answer": "2"}, {"score": 0.296, "answer": "1"}]
)
outputs = vqa_pipeline({"image": image, "question": question}, top_k=2)
self.assertEqual(
nested_simplify(outputs, decimals=4), [{"score": 0.8799, "answer": "2"}, {"score": 0.296, "answer": "1"}]
)
outputs = vqa_pipeline(
[{"image": image, "question": question}, {"image": image, "question": question}], top_k=2
)
self.assertEqual(
nested_simplify(outputs, decimals=4),
[[{"score": 0.8799, "answer": "2"}, {"score": 0.296, "answer": "1"}]] * 2,
)
@slow
@require_torch
@require_torch_gpu
def test_large_model_pt_blip2(self):
vqa_pipeline = pipeline(
"visual-question-answering",
model="Salesforce/blip2-opt-2.7b",
model_kwargs={"torch_dtype": torch.float16},
device=0,
)
self.assertEqual(vqa_pipeline.model.device, torch.device(0))
self.assertEqual(vqa_pipeline.model.language_model.dtype, torch.float16)
image = "./tests/fixtures/tests_samples/COCO/000000039769.png"
question = "Question: how many cats are there? Answer:"
outputs = vqa_pipeline(image=image, question=question)
self.assertEqual(outputs, [{"answer": "two"}])
outputs = vqa_pipeline({"image": image, "question": question})
self.assertEqual(outputs, [{"answer": "two"}])
outputs = vqa_pipeline([{"image": image, "question": question}, {"image": image, "question": question}])
self.assertEqual(outputs, [[{"answer": "two"}]] * 2)
@require_tf
@unittest.skip("Visual question answering not implemented in TF")
def test_small_model_tf(self):
pass
|