Instructions to use David-Magdy/TR_OCR_LARGE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use David-Magdy/TR_OCR_LARGE with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="David-Magdy/TR_OCR_LARGE")

# Load model directly
from transformers import AutoTokenizer, AutoModelForImageTextToText

tokenizer = AutoTokenizer.from_pretrained("David-Magdy/TR_OCR_LARGE")
model = AutoModelForImageTextToText.from_pretrained("David-Magdy/TR_OCR_LARGE")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use David-Magdy/TR_OCR_LARGE with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "David-Magdy/TR_OCR_LARGE"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "David-Magdy/TR_OCR_LARGE",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/David-Magdy/TR_OCR_LARGE

SGLang

How to use David-Magdy/TR_OCR_LARGE with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "David-Magdy/TR_OCR_LARGE" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "David-Magdy/TR_OCR_LARGE",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "David-Magdy/TR_OCR_LARGE" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "David-Magdy/TR_OCR_LARGE",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use David-Magdy/TR_OCR_LARGE with Docker Model Runner:
```
docker model run hf.co/David-Magdy/TR_OCR_LARGE
```

library_name: transformers tags:

OCR
handwritten-text-recognition
multilingual
Arabic
English

Model Card for TR-OCR Large AR/EN Handwritten

This is a finetuned version of TROCR Large specialized in handwritten text recognition for Arabic and English languages.

Model Details

Model Description

This is a finetuned version of Microsoft's TROCR Large model, adapted for handwritten text recognition in Arabic and English languages using the Khatt and IAM Handwriting datasets.

Developed by: Me and my colleague Ahmed Wahdan
Model type: OCR (Optical Character Recognition)
Language(s) (NLP): Arabic, English
Finetuned from model: Microsoft TROCR Large

Model Sources [optional]

Repository: Kaggle Notebook - Yet to be provided
Original Model Paper: TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models

Uses

Direct Use

This model is intended for handwritten text recognition in Arabic and English documents.

Out-of-Scope Use

The model should not be used for:

Languages other than Arabic and English
Printed text recognition
Non-text image analysis

Bias, Risks, and Limitations

Limitations

Only supports Arabic and English languages
Performance may vary with different handwriting styles
Not tested on all possible handwriting variations

Recommendations

Users should be aware that the model is specifically trained for Arabic and English handwritten text and may not perform well on other languages or printed text.

How to Get Started with the Model

# Sample code to load the model
from transformers import TrOCRProcessor, VisionEncoderDecoderModel

processor = TrOCRProcessor.from_pretrained("David-Magdy/TR_OCR_LARGE")
model = VisionEncoderDecoderModel.from_pretrained("David-Magdy/TR_OCR_LARGE")

Downloads last month: 445

Safetensors

Model size

0.6B params

Tensor type

F32

Inference Providers NEW

Image-Text-to-Text

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for David-Magdy/TR_OCR_LARGE

Base model

microsoft/trocr-large-handwritten

Finetuned

(15)

this model

Dataset used to train David-Magdy/TR_OCR_LARGE

Paper for David-Magdy/TR_OCR_LARGE

TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models

Paper • 2109.10282 • Published Sep 21, 2021 • 13