-
How to convert to ONNX? I want to convert PaddleOCR v4 server to ONNX myself, but after using the built-in PaddleOCR to convert to ONNX, the decoding of /ocr/PaddleOCR/ppocr/utils/ppocr_keys_v1.txt re…
xddun updated
1 month ago
-
**Describe the bug**
instead of "succeeded" it should read "FAILED"
[workflow_ocr] Warning: OCRmyPDF FAILED with warning(s): sh: ocrmypdf: not found
this BTW happens because these packages are …
-
**Describe the bug**
Sending a single PDF (this one: https://arxiv.org/abs/2310.12931, embedded text, 39 pages) to the self hosted API, with either ocr_only or hi_res strategy.
The server has 32GB R…
-
# MindSpore OCR 服务化部署功能设计说明书
## 一、修订记录
| ***\*日期\**** | ***\*修订版本\**** | ***\*修改章节\**** | ***\*修改描述\**** | ***\*作者\**** |
| -------------- | ------------------ | ------------------ | ---…
-
I was able to run the application after your latest update. Hurray!! Now, I am unable to use the OCR and STT. Both show the "Failed to reach OCR server. Is the server running?" message.
![OCR message…
-
## Describe the bug
Trying to upload a new version of dataset (the first version has already been downloaded), but getting the error (never happened before):
```
2024-05-06 14:21:37,529 - clearml.T…
-
## 🚀 Feature
Multiple endpoints like `/embedding` or `/vlm/predict` or `/ocr/predict`.
### Motivation
I would like to host multiple models on a single GPU for different purposes. It would be …
-
![Screenshot 2024-08-24 160058](https://github.com/user-attachments/assets/d5e6993a-982d-4377-b9a7-2698a50a1340)
Debug log shows "non empty source txt list"
OG comic language is Dutch (Netherlan…
-
Would you consider including OCR in the mobile app (or even better on the server end). So scanned or photographed invoices (expenses) can be populated with data automatically and saved as drafts (with…
-
1、在百度网盘下载好了这个100多兆的rec.onnx文件
2、然后放在了onnxocr/models/ch_ppocr_server_v2.0/rec文件夹下,
3、在utils.py文件的def infer_args():中修改了配置为:
parser.add_argument("--det_model_dir", type=str, default='./onnxocr/m…