-
非常感谢badboy的工作，👍赞
记录下自己在使用项目过程中遇见的问题
1、
问题：torch 用不了cuda
原因：pytorch==1.5.0 是用pip安装的，导致torch和cuda可能不匹配，具体匹配规则可见https://blog.csdn.net/kellyroslyn/article/details/109668001
解决：去torch官网下载https://pyt…
-
您好，我还没有开始训练自己的模型，正在用预训练的模型跑demo.py,我发现速度特别的慢，5分钟左右demo.py才能跑出结果，这个模型本来就慢吗?还是我自身的问题
-
-
Not sure if this is the right place to post, but posting here since Catalyst seems only place related to Chemical reactions. I recently across paper deriving unknown chemical reaction pathways using n…
-
I have tried to load the table transformer detection and segmentation model in the doctection analyzer as indicated in the repo but I get the following error:
Model not found in ModelCatalog. Make s…
-
hello,have you tried use more alphabet?
I have followed your ways to trianed rcnn model with just 10 numbers,10 chinese chars,10 mixed chinese chars and numbers and lower alpha,these model all perfor…
-
### Prerequisite
- [X] I have searched [Issues](https://github.com/open-mmlab/mmocr/issues) and [Discussions](https://github.com/open-mmlab/mmocr/discussions) but cannot get the expected help.
- […
-
CUDA_VISIBLE_DEVICES=GPU_NUM python main.py --batch_size=16 --STN --exp_name EXP_NAME --text_focus --resume YOUR_MODEL --test --test_data_dir ./dataset/mydata/test
作者您好，请问这个YOUR_MODEL 具体应该填啥，我试了很多次都没…
-
paddleocr模型能转到1.6.0版本的onnx，且精度没有问题，但1.6.0版本的onnx转rknn遇到了单向双层的lstm算子不支持问题，即使把onnx的选项opset-version设置成11/12也是报出了同样的问题，
```text
W Not match tensor LSTM_LSTM_4:out0
E Try match layer: LSTM_LSTM_4:out0 …
-
I am new to OCR. I run your code and the Loss and Recognition Loss are all negative and detection loss is positive. I am a little confused about this result. Is it correct?
Train Epoch: 3 [576/856 …