Upload folder using huggingface_hub
Browse files- .gitattributes +1 -1
- README.md +17 -2
- README_CN.md +15 -3
- vis.gif +3 -0
.gitattributes
CHANGED
@@ -64,4 +64,4 @@ resources/fonts/方正宋黑.TTF filter=lfs diff=lfs merge=lfs -text
|
|
64 |
resources/fonts/japan.ttc filter=lfs diff=lfs merge=lfs -text
|
65 |
resources/fonts/FZYTK.TTF filter=lfs diff=lfs merge=lfs -text
|
66 |
resources/fonts/korean.ttf filter=lfs diff=lfs merge=lfs -text
|
67 |
-
resources/fonts/方正宋黑.TTF filter=lfs diff=lfs merge=lfs -text
|
|
|
64 |
resources/fonts/japan.ttc filter=lfs diff=lfs merge=lfs -text
|
65 |
resources/fonts/FZYTK.TTF filter=lfs diff=lfs merge=lfs -text
|
66 |
resources/fonts/korean.ttf filter=lfs diff=lfs merge=lfs -text
|
67 |
+
resources/fonts/方正宋黑.TTF filter=lfs diff=lfs merge=lfs -textvis.gif filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
@@ -31,11 +31,23 @@ Current token-prediction-based model architectures are highly sensitive to the a
|
|
31 |
|
32 |
## Visualization
|
33 |
|
34 |
-

|
35 |
-
|
36 |
## Installation
|
37 |
|
38 |
```bash
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
pip install phocr
|
40 |
```
|
41 |
|
@@ -54,6 +66,9 @@ print(result)
|
|
54 |
# Visualize results
|
55 |
result.vis("output.jpg")
|
56 |
print(result.to_markdown())
|
|
|
|
|
|
|
57 |
```
|
58 |
|
59 |
## Benchmarks
|
|
|
31 |
|
32 |
## Visualization
|
33 |
|
|
|
|
|
34 |
## Installation
|
35 |
|
36 |
```bash
|
37 |
+
# Choose **one** installation method below:
|
38 |
+
|
39 |
+
# Method 1: Install with ONNX Runtime CPU version
|
40 |
+
pip install phocr[cpu]
|
41 |
+
|
42 |
+
# Method 2: Install with ONNX Runtime GPU version
|
43 |
+
pip install phocr[cuda]
|
44 |
+
# Required: Make sure the CUDA toolkit and cuDNN library are properly installed
|
45 |
+
# You can install cuda runtime and cuDNN via conda:
|
46 |
+
conda install -c nvidia cuda-runtime=12.1 cudnn=9
|
47 |
+
# Or manually install the corresponding CUDA toolkit and cuDNN libraries
|
48 |
+
|
49 |
+
# Method 3: Manually manage ONNX Runtime
|
50 |
+
# You can install `onnxruntime` or `onnxruntime-gpu` yourself, then install PHOCR
|
51 |
pip install phocr
|
52 |
```
|
53 |
|
|
|
66 |
# Visualize results
|
67 |
result.vis("output.jpg")
|
68 |
print(result.to_markdown())
|
69 |
+
|
70 |
+
## only recognition
|
71 |
+
|
72 |
```
|
73 |
|
74 |
## Benchmarks
|
README_CN.md
CHANGED
@@ -32,11 +32,22 @@ PHOCR 是一个高性能的开源光学字符识别(OCR)工具包,专为
|
|
32 |
|
33 |
## 可视化效果
|
34 |
|
35 |
-

|
36 |
-
|
37 |
## 安装方式
|
38 |
-
|
39 |
```bash
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
pip install phocr
|
41 |
```
|
42 |
|
@@ -55,6 +66,7 @@ print(result)
|
|
55 |
# 可视化结果
|
56 |
result.vis("output.jpg")
|
57 |
print(result.to_markdown())
|
|
|
58 |
```
|
59 |
|
60 |
## 性能基准测试
|
|
|
32 |
|
33 |
## 可视化效果
|
34 |
|
|
|
|
|
35 |
## 安装方式
|
|
|
36 |
```bash
|
37 |
+
# 请选择以下 **一种** 安装方式:
|
38 |
+
|
39 |
+
# 方法一:安装带 ONNX Runtime CPU 版本
|
40 |
+
pip install phocr[cpu]
|
41 |
+
|
42 |
+
# 方法二:安装带 ONNX Runtime GPU 版本
|
43 |
+
pip install phocr[cuda]
|
44 |
+
# 必须:确保已正确安装 CUDA Toolkit 和 cuDNN 库
|
45 |
+
# 你可以通过 conda 安装 cuda-runtime和cuDNN:
|
46 |
+
conda install -c nvidia cuda-runtime=12.1 cudnn=9
|
47 |
+
# 或者手动安装对应版本的 CUDA Toolkit 和 cuDNN 库
|
48 |
+
|
49 |
+
# 方法三:手动管理 ONNX Runtime
|
50 |
+
# 你可以自行安装 `onnxruntime` 或 `onnxruntime-gpu`,然后再安装 PHOCR
|
51 |
pip install phocr
|
52 |
```
|
53 |
|
|
|
66 |
# 可视化结果
|
67 |
result.vis("output.jpg")
|
68 |
print(result.to_markdown())
|
69 |
+
|
70 |
```
|
71 |
|
72 |
## 性能基准测试
|
vis.gif
ADDED
![]() |
Git LFS Details
|