mirror of
https://github.com/modelscope/modelscope.git
synced 2025-12-25 20:49:37 +01:00
cba4e40bc1
明确受影响的模型(damo)： ONE-PEACE-4B ModuleNotFoundError: MyCustomPipeline: MyCustomModel: No module named 'one_peace'，缺少依赖。 cv_resnet50_face-reconstruction 不兼容tf2 nlp_automatic_post_editing_for_translation_en2de tf2.0兼容性问题，tf1.x需要 cv_resnet18_ocr-detection-word-level_damo tf2.x兼容性问题 cv_resnet18_ocr-detection-line-level_damo tf兼容性问题 cv_resnet101_detection_fewshot-defrcn 模型限制必须detection0.3+torch1.11.0" speech_dfsmn_ans_psm_48k_causal "librosa， numpy兼容性问题 cv_mdm_motion-generation "依赖numpy版本兼容性问题： File ""/opt/conda/lib/python3.8/site-packages/smplx/body_models.py"", cv_resnet50_ocr-detection-vlpt numpy兼容性问题 cv_clip-it_video-summarization_language-guided_en tf兼容性问题 Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13744636 * numpy and pandas no version * modify compatible issue * fix numpy compatible issue * modify ci * fix lint issue * replace Image.ANTIALIAS to Image.Resampling.LANCZOS pillow compatible * skip uncompatible cases * fix numpy compatible issue, skip cases that can not compatbile numpy or tensorflow2.x * skip compatible cases * fix clip model issue * fix body 3d keypoints compatible issue
99 lines
3.4 KiB
Python
99 lines
3.4 KiB
Python
# Copyright (c) Alibaba, Inc. and its affiliates.
|
|
import os
|
|
import shutil
|
|
import tempfile
|
|
import unittest
|
|
|
|
from modelscope.hub.snapshot_download import snapshot_download
|
|
from modelscope.metainfo import Trainers
|
|
from modelscope.models.cv.ocr_recognition import OCRRecognition
|
|
from modelscope.msdatasets import MsDataset
|
|
from modelscope.trainers import build_trainer
|
|
from modelscope.utils.config import Config, ConfigDict
|
|
from modelscope.utils.constant import DownloadMode, ModelFile
|
|
from modelscope.utils.test_utils import test_level
|
|
|
|
|
|
@unittest.skip(
|
|
"For FileNotFoundError: [Errno 2] No such file or directory: './work_dir/output/pytorch_model.pt' issue"
|
|
)
|
|
class TestOCRRecognitionTrainer(unittest.TestCase):
|
|
|
|
model_id = 'damo/cv_crnn_ocr-recognition-general_damo'
|
|
|
|
def setUp(self):
|
|
print(('Testing %s.%s' % (type(self).__name__, self._testMethodName)))
|
|
|
|
cache_path = snapshot_download(self.model_id, revision='v2.2.2')
|
|
config_path = os.path.join(cache_path, ModelFile.CONFIGURATION)
|
|
cfg = Config.from_file(config_path)
|
|
|
|
max_epochs = cfg.train.max_epochs
|
|
|
|
train_data_cfg = ConfigDict(
|
|
name='ICDAR13_HCTR_Dataset', split='test', namespace='damo')
|
|
|
|
test_data_cfg = ConfigDict(
|
|
name='ICDAR13_HCTR_Dataset', split='test', namespace='damo')
|
|
|
|
self.train_dataset = MsDataset.load(
|
|
dataset_name=train_data_cfg.name,
|
|
split=train_data_cfg.split,
|
|
namespace=train_data_cfg.namespace,
|
|
download_mode=DownloadMode.REUSE_DATASET_IF_EXISTS)
|
|
assert next(
|
|
iter(self.train_dataset.config_kwargs['split_config'].values()))
|
|
|
|
self.test_dataset = MsDataset.load(
|
|
dataset_name=test_data_cfg.name,
|
|
split=test_data_cfg.split,
|
|
namespace=train_data_cfg.namespace,
|
|
download_mode=DownloadMode.REUSE_DATASET_IF_EXISTS)
|
|
assert next(
|
|
iter(self.test_dataset.config_kwargs['split_config'].values()))
|
|
|
|
self.max_epochs = max_epochs
|
|
|
|
self.tmp_dir = tempfile.TemporaryDirectory().name
|
|
if not os.path.exists(self.tmp_dir):
|
|
os.makedirs(self.tmp_dir)
|
|
|
|
def tearDown(self):
|
|
shutil.rmtree(self.tmp_dir)
|
|
super().tearDown()
|
|
|
|
@unittest.skipUnless(test_level() >= 0, 'skip test in current test level')
|
|
def test_trainer(self):
|
|
kwargs = dict(
|
|
model=self.model_id,
|
|
train_dataset=self.train_dataset,
|
|
eval_dataset=self.test_dataset,
|
|
work_dir=self.tmp_dir)
|
|
|
|
trainer = build_trainer(
|
|
name=Trainers.ocr_recognition, default_args=kwargs)
|
|
trainer.train()
|
|
|
|
@unittest.skipUnless(test_level() >= 0, 'skip test in current test level')
|
|
def test_trainer_with_model_and_args(self):
|
|
tmp_dir = tempfile.TemporaryDirectory().name
|
|
if not os.path.exists(tmp_dir):
|
|
os.makedirs(tmp_dir)
|
|
|
|
cache_path = snapshot_download(self.model_id, revision='v2.2.2')
|
|
model = OCRRecognition.from_pretrained(cache_path)
|
|
kwargs = dict(
|
|
cfg_file=os.path.join(cache_path, ModelFile.CONFIGURATION),
|
|
model=model,
|
|
train_dataset=self.train_dataset,
|
|
eval_dataset=self.test_dataset,
|
|
work_dir=tmp_dir)
|
|
|
|
trainer = build_trainer(
|
|
name=Trainers.ocr_recognition, default_args=kwargs)
|
|
trainer.train()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
unittest.main()
|