Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)
[2024-03-08 09:55:08,197] [ INFO] - All the weights of LayoutXLMModel were initialized from the model checkpoint at vi-layoutxlm-base-uncased.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LayoutXLMModel for predictions without further training.
Traceback (most recent call last):
File "tools/train.py", line 227, in
main(config, device, logger, vdl_writer)
File "tools/train.py", line 135, in main
model = build_model(config['Architecture'])
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/init.py", line 34, in build_model
arch = getattr(mod, name)(config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/distillation_model.py", line 47, in init
model = BaseModel(model_config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/base_model.py", line 55, in init
self.backbone = build_backbone(config["Backbone"], model_type)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/init.py", line 80, in build_backbone
module_class = eval(module_name)(*config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/vqa_layoutlm.py", line 149, in init
super(LayoutXLMForSer, self).init(
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/vqa_layoutlm.py", line 65, in init
self.model = model_class(
File "/Users/zhuxiaoxia/.pyenv/versions/3.8.12/lib/python3.8/site-packages/paddlenlp/transformers/utils.py", line 253, in impl
init_func(self, args, **kwargs)
TypeError: init() got an unexpected keyword argument 'num_classes'
➜ PaddleOCR git:(release/2.7) ✗
环境如下
[2024-03-08 09:55:08,197] [ INFO] - All the weights of LayoutXLMModel were initialized from the model checkpoint at vi-layoutxlm-base-uncased. If your task is similar to the task the model of the checkpoint was trained on, you can already use LayoutXLMModel for predictions without further training. Traceback (most recent call last): File "tools/train.py", line 227, in
main(config, device, logger, vdl_writer)
File "tools/train.py", line 135, in main
model = build_model(config['Architecture'])
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/init.py", line 34, in build_model
arch = getattr(mod, name)(config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/distillation_model.py", line 47, in init
model = BaseModel(model_config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/architectures/base_model.py", line 55, in init
self.backbone = build_backbone(config["Backbone"], model_type)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/init.py", line 80, in build_backbone
module_class = eval(module_name)(*config)
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/vqa_layoutlm.py", line 149, in init
super(LayoutXLMForSer, self).init(
File "/Users/zhuxiaoxia/Documents/python/PaddleOCR/ppocr/modeling/backbones/vqa_layoutlm.py", line 65, in init
self.model = model_class(
File "/Users/zhuxiaoxia/.pyenv/versions/3.8.12/lib/python3.8/site-packages/paddlenlp/transformers/utils.py", line 253, in impl
init_func(self, args, **kwargs)
TypeError: init() got an unexpected keyword argument 'num_classes'
➜ PaddleOCR git:(release/2.7) ✗