-
根据readme config.json
```
{
"modules_info": {
"chinese_ocr_db_crnn_mobile": {
"init_args": {
"version": "1.0.3"
},
"predict_args": {
"use_gpu": true,
…
BobDu updated
4 years ago
-
paddlehub==1.8.3
paddlepaddle==1.8.5
Shapely==1.7.1
pyclipper==1.2.0
numpy==1.19.1
![0](https://user-images.githubusercontent.com/14155359/100200305-5889cf80-2f39-11eb-98e2-f6799d867898.png)
-
Thank you for contributing to PaddlePaddle.
Before submitting the issue, you could search issue in the github in case that th
If there is no solution,please make sure that this is an inference issue…
-
我现在使用Flask自己起一个服务,把ch_det_r50_vd_db和ch_rec_r34_vd_crnn 2个模型初始化常驻在内存,本身这2个模型模型文件也就200M左右,但是常驻内存占了大概有7G的内存,请问正常嘛?有没什么办法释放,常驻的原因是加快推理速度,不用每次请求都重新初始化模型
-
## 资源
- OCR dataset [[github]](https://github.com/WenmuZhou/OCR_DataSet)
- awesome-deep-text-detection-recognition [[github]](https://github.com/hwalsuklee/awesome-deep-text-detection-recognition)…
-
![image](https://user-images.githubusercontent.com/46363468/88631110-36e01f00-d0e4-11ea-9e61-098bce856779.png)
-
没法下载安装模型
D:\Programs\PaddleOCR>hub install chinese_ocr_db_crnn_mobile
Can't find module chinese_ocr_db_crnn_mobile, please check your spelling.
但下面这个识图可以装成功~
hub install deploy\hubserving\ocr_sy…
-
在使用超轻量模型时可以正常得到OCR信息,但当我对同一张图片使用通用中文模型时会报如下错误
```
λ iZwz9e694cocz8fn3et9d9Z /home/PaddleOCR {develop} python3 tools/infer/predict_system.py --image_dir="/opt/tmp/fb_1100.jpg" --det_model_dir="infere…
-
I used the pytorch code [crnn](https://github.com/meijieru/crnn.pytorch/blob/master/models/crnn.py) and the cnn backbone change as mobilenet v3 large 1.0, to train a ocr recognition model, the input …
DuckJ updated
4 years ago
-
硬件:Tesla P100。
YOLO检测速度很快,但是crnn的识别感觉略慢。如下所示,光识别所有行需要2s左右。而我用另外的densenet模型识别只需要0.45s. 如下图所示(ocr_batch execution time):
请问crnn这个速度正常嘛?大概优化点在哪里 谢谢。
CRNN:
======
Densenet:
打印出的数据是概率…