KeyError: 'pytorch-lightning_version'

[laster-lee]

Traceback (most recent call last): File "e:\kc_infer\kc_test.py", line 7, in parseq = load_from_checkpoint("./parseq_tiny.pt").eval() File "e:\kc_infer\strhub\models\utils.py", line 92, in load_from_checkpoint model = ModelClass.load_from_checkpoint(checkpoint_path, *kwargs) File "D:\anaconda3\envs\kc_infer\lib\site-packages\pytorch_lightning\utilities\model_helpers.py", line 125, in wrapper return self.method(cls, args, **kwargs) File "D:\anaconda3\envs\kc_infer\lib\site-packages\pytorch_lightning\core\module.py", line 1581, in load_from_checkpoint loaded = _load_from_checkpoint( File "D:\anaconda3\envs\kc_infer\lib\site-packages\pytorch_lightning\core\saving.py", line 66, in _load_from_checkpoint checkpoint = _pl_migrate_checkpoint( File "D:\anaconda3\envs\kc_infer\lib\site-packages\pytorch_lightning\utilities\migration\utils.py", line 143, in _pl_migrate_checkpoint old_version = _get_version(checkpoint) File "D:\anaconda3\envs\kc_infer\lib\site-packages\pytorch_lightning\utilities\migration\utils.py", line 164, in _get_version return checkpoint["pytorch-lightning_version"] KeyError: 'pytorch-lightning_version'

[EtherealTan]

i have the same question have you solved it?

[laster-lee]

i have the same question have you solved it?我有同样的问题，你解决了吗？

没解决，后面我自己训练模型了，用自己的数据集

[GIN-Margarita]

you need execuate with --checkpoint pretrained=

example: python read.py --checkpoint pretrained=vitstr --images ./demo_images/art-01107.jpg ./demo_images/coco-1166773.jpg meanwhile ,you need update read.py args with --checkpoints , such as :

[GIN-Margarita]

additionally , utils.py has urls of model you could update url with the local dir.