trtr
#16
by
sdsd456457541
- opened
README.md
CHANGED
|
@@ -30,6 +30,20 @@ Images are presented to the model as a sequence of fixed-size patches (resolutio
|
|
| 30 |
|
| 31 |
You can use the raw model for optical character recognition (OCR) on single text-line images. See the [model hub](https://huggingface.co/models?search=microsoft/trocr) to look for fine-tuned versions on a task that interests you.
|
| 32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
### How to use
|
| 34 |
|
| 35 |
Here is how to use this model in PyTorch:
|
|
|
|
| 30 |
|
| 31 |
You can use the raw model for optical character recognition (OCR) on single text-line images. See the [model hub](https://huggingface.co/models?search=microsoft/trocr) to look for fine-tuned versions on a task that interests you.
|
| 32 |
|
| 33 |
+
### Performance Characteristics
|
| 34 |
+
|
| 35 |
+
This model works best on:
|
| 36 |
+
- Single-line handwritten text
|
| 37 |
+
- Clear, well-lit images
|
| 38 |
+
- Text in English language
|
| 39 |
+
- Images with good contrast between text and background
|
| 40 |
+
|
| 41 |
+
### Limitations
|
| 42 |
+
|
| 43 |
+
- May struggle with multi-line text detection
|
| 44 |
+
- Performance may degrade on cursive or heavily stylized handwriting
|
| 45 |
+
- Not optimized for printed text (consider other TrOCR variants for printed text)
|
| 46 |
+
|
| 47 |
### How to use
|
| 48 |
|
| 49 |
Here is how to use this model in PyTorch:
|