DeepSolo
DeepSolo copied to clipboard
ViTAE-Transformer
The official repo for [CVPR'23] "DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting" & [ArXiv'23] "DeepSolo++: Let Transformer Decoder with Explicit Points Solo for Multilin...
Traceback (most recent call last): File "tools/train_net.py", line 43, in from adet.data.dataset_mapper import DatasetMapperWithBasis File "/root/autodl-tmp/DeepSolo-main/adet/__init__.py", line 1, in from adet import modeling File "/root/autodl-tmp/DeepSolo-main/adet/modeling/__init__.py", line 2, in from .text_spotter...
Hello, thanks for your great work. I wish train a model on my own dataset, yet I don't know how to generate them for I don't know how te get...
thks your proj! how can i use the tensorrt to inference my custom trained model. I want to deploy the model in Jetson series，thks for your reply。
 Hi there! I was really impressed by the visualization results in the paper. I managed to find the code for visualizing spotting results in the repository, which was helpful....
This PR implements DeepSolo on CPU by calling the pytorch implementation of multiscale deformable attention if cuda is not found. I have also organised the cpu/cuda files into cpu/cuda folders...
I want to create my custom dataset for training with DeepSolo. But I don't know how to do it. I know there is a tool called PPOCRLabel provided by PaddleOCR...
我添加了： ``` os.environ["CUDA_VISIBLE_DEVICES"] = "1,2" ``` 其中，指定 ``` "args": [ "--config-file", "configs/FPN_R_50/CTW1500/finetune_96voc_50maxlen.yaml", "--num-gpus", "2" ], ``` 在train_net.py中，并且torch.cuda.device_count() = 2 报错： ``` File "/home/wit/anaconda3/envs/lj_py3.8pt1.10/lib/python3.8/site-packages/detectron2/engine/launch.py", line 119, in _distributed_worker assert num_gpus_per_machine
No such file or directory: 'datasets/totaltext/train_37voc.json'，Need to modify train.json to train_37voc.json?
Metadata
Owner
Metadata
The official repo for [CVPR'23] "DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting" & [ArXiv'23] "DeepSolo++: Let Transformer Decoder with Explicit Points Solo for Multilin...