ModelListedExternal

Qari OCR 0.4.0 VL 4B Instruct

A vision-language model fine-tuned for OCR on Islamic books and Arabic manuscripts. Based on Qwen/Qwen3-VL-4B-Instruct, trained on 45,000 image-text pairs from the seemorg/books-ocr dataset.

Sourcehuggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct/tree/aa65b4a46336c6cc608e7da5a015f8e71472a68b

KindLoRABase modelQwen/Qwen3-VL-4B-InstructVersionadapter_model

Model source

Kind: LoRA
File size: 300 MB
Base model: Qwen/Qwen3-VL-4B-Instruct
Version: vadapter_model
Updated: Sep 26, 2026
Downloads: 734
Tasks: Captioning
Source: Hugging Face

Source description

Qari-OCR-0.4.0-VL-4B-Instruct

A vision-language model fine-tuned for OCR on Islamic books and Arabic manuscripts. Based on Qwen/Qwen3-VL-4B-Instruct, trained on 45,000 image-text pairs from the seemorg/books-ocr dataset.

Results

Model	CER ↓	WER ↓	BLEU ↑
Qari-OCR-0.4.0	0.1222	0.2562	68.41
Qwen/Qwen3-VL-4B-Instruct	0.4922	0.6966	34.61
Qwen/Qwen3-VL-8B-Instruct	0.6876	0.8954	23.89
NAMAA/Qari-0.2.2.1	0.6448	0.5126	21.97
MBZUAI/AIN	1.2843	1.2697	3.50

Usage

from transformers import AutoProcessor, AutoModelForVision2Seq
from PIL import Image
import torch

model_name = "NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct"
processor = AutoProcessor.from_pretrained(model_name)
model = AutoModelForVision2Seq.from_pretrained(
    model_name,
    torch_dtype=torch.float16,
    device_map="auto"
)

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": f"./{src}"},
            {"type": "text", "text":  "Free OCR."},
        ],
    }
]

# Preparation for inference
inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_dict=True,
    return_tensors="pt"
)
inputs = inputs.to(model.device)

# Inference: Generation of the output
generated_ids = model.generate(**inputs, max_new_tokens=2048)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
result = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)[0]

Training

Base model: Qwen/Qwen3-VL-4B-Instruct
Dataset: seemorg/books-ocr
Training samples: 45,000 image-text pairs
Domain: Islamic books and Arabic religious texts

Limitations

Optimized for printed Islamic texts; performance may vary on modern Arabic fonts or handwritten text.
Requires reasonable image quality (300+ DPI recommended).
Arabic script only.

Citation

@misc{qari-ocr-0.4.0,
  author       = {NAMAA-Space},
  title        = {Qari-OCR-0.4.0-VL-4B-Instruct},
  year         = {2025},
  publisher    = {HuggingFace},
  howpublished = {\url{https://huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct}}
}

Sources

1 source

Hugging FacePrimaryVersion aa65b4a46336c6cc608e7da5a015f8e71472a68b

huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct/tree/aa65b4a46336c6cc608e7da5a015f8e71472a68b

Verified Sep 11

Model artifacts

1 artifact

adapter_model.safetensors

safetensors · 300 MB · SHA-256 892c3601aa35…e99a · Hugging Face

Download

Source excerpts

2 excerpts

ReadmeSep 11huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct

--- license: apache-2.0 base_model: Qwen/Qwen3-VL-4B-Instruct tags: - vision - ocr - arabic - islamic - qwen3 - fine-tuned language: - ar datasets: - seemorg/books-ocr metrics: - wer - cer - bleu library_name: peft pipeline_tag: image-to-text --- # Qari-OCR-0.4.0-VL-4B-Instruct A vision-language model fine-tuned for OCR on Islamic books and Arabic manuscripts. Based on [Qwen/Qwen3-VL-4B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct), trained on 45,000 image-text pairs from the [seemorg/books-ocr](https://huggingface.co/datasets/seemorg/books-ocr) dataset. ## Results | Model | CER ↓ | WER ↓ | BLEU ↑ | |-------|-------|-------|--------| | **Qari-OCR-0.4.0** | **0.1222** | **0.2562** | **68.41** | | Qwen/Qwen3-VL-4B-Instruct | 0.4922 | 0.6966 | 34.61 | | Qwen/Qwen3-VL-8B-Instruct | 0.6876 | 0.8954 | 23.89 | | NAMAA/Qari-0.2.2.1 | 0.6448 | 0.5126 | 21.97 | | MBZUAI/AIN | 1.2843 | 1.2697 | 3.50 | ## Usage ```python from transformers import AutoProcessor, AutoModelForVision2Seq from PIL import Image import torch model_name = "NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct" processor = AutoProcessor.from_pretrained(model_name) model = AutoModelForVision2Seq.from_pretrained( model_name, torch_dtype=torch.float16, device_map="auto" ) messages = [ { "role": "user", "content": [ {"type": "image", "image": f"./{src}"}, {"type": "text", "text": "Free OCR."}, ], } ] # Preparation for inference inputs = processor.apply_chat_template( messages, tokenize=True, add_generation_prompt=True, return_dict=True, return_tensors="pt" ) inputs = inputs.to(model.device) # Inference: Generation of the output generated_ids = model.generate(**inputs, max_new_tokens=2048) generated_ids_trimmed = [ out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids) ] result = processor.batch_decode( generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False )[0] ``` ## Training - **Base model:** Qwen/Qwen3-VL-4B-Instruct - **Dataset:** [seemorg/books-ocr](https://huggingface.co/datasets/seemorg/books-ocr) - **Training samples:** 45,000 image-text pairs - **Domain:** Islamic books and Arabic religious texts ## Limitations - Optimized for printed Islamic texts; performance may vary on modern Arabic fonts or handwritten text. - Requires reasonable image quality (300+ DPI recommended). - Arabic script only. ## Ci...

NoteSep 11huggingface.co/NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct

Source context: 734 downloads · 7 likes · Pipeline image-to-text · Library peft · Repo NAMAA-Space/Qari-OCR-0.4.0-VL-4B-Instruct