suchut commited on
Commit
db95d39
1 Parent(s): 6d59796

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +61 -0
README.md ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - th
4
+ - en
5
+ metrics:
6
+ - cer
7
+ tags:
8
+ - trocr
9
+ - image-to-text
10
+ pipeline_tag: image-to-text
11
+ library_name: transformers
12
+ license: apache-2.0
13
+ ---
14
+ # Thai-TrOCR Model
15
+
16
+ ## 🚀 Final Model Available Now!
17
+ **The final version of the Thai-TrOCR model is out!** Check it out here: [huggingface.com/openthaigpt/thai-trocr](https://huggingface.co/openthaigpt/thai-trocr)
18
+
19
+ ---
20
+
21
+ ## Introduction
22
+ **Thai-TrOCR** is an advanced Optical Character Recognition (OCR) model fine-tuned specifically for recognizing handwritten text in **Thai** and **English**. Built on the robust TrOCR architecture, this model combines a Vision Transformer encoder with an Electra-based text decoder, allowing it to effectively handle multilingual text-line images.
23
+
24
+ Designed for **efficiency and accuracy**, Thai-TrOCR is lightweight, making it ideal for deployment in resource-constrained environments without compromising on performance.
25
+
26
+ ### Key Features:
27
+ - **Encoder**: TrOCR Base Handwritten
28
+ - **Decoder**: Electra Small (Trained with Thai corpus)
29
+
30
+ ---
31
+
32
+ ## Training Dataset
33
+ Thai-TrOCR was trained using the following datasets:
34
+ - `pythainlp/thai-wiki-dataset-v3`
35
+ - `pythainlp/thaigov-corpus`
36
+ - `Salesforce/wikitext`
37
+
38
+ ---
39
+
40
+ ## How to Use This Beta Model
41
+ Here’s a quick guide to get started with the Thai-TrOCR model in **PyTorch**:
42
+
43
+ ```python
44
+ from transformers import TrOCRProcessor, VisionEncoderDecoderModel
45
+ from PIL import Image
46
+ import requests
47
+
48
+ # Load processor and model
49
+ processor = TrOCRProcessor.from_pretrained('suchut/thaitrocr-base-handwritten-beta3')
50
+ model = VisionEncoderDecoderModel.from_pretrained('suchut/thaitrocr-base-handwritten-beta3')
51
+
52
+ # Load an image
53
+ url = 'your_image_url_here'
54
+ image = Image.open(requests.get(url, stream=True).raw).convert("RGB")
55
+
56
+ # Process and generate text
57
+ pixel_values = processor(images=image, return_tensors="pt").pixel_values
58
+ generated_ids = model.generate(pixel_values)
59
+ generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
60
+ print(generated_text)
61
+ ```