---
library_name: transformers
tags:
- thaana-ocr
- dhivehi-ocr
license: mit
datasets:
- alakxender/dhivehi-vrd-images
language:
- dv
base_model:
- unsloth/DeepSeek-OCR
---

#  DeepSeek-OCR Dhivehi

This model is a fine-tuned version of [**unsloth/DeepSeek-OCR**](https://huggingface.co/unsloth/DeepSeek-OCR), trained for **Dhivehi single-line sentence recognition**.
It was fine-tuned using **50,000 samples** from the [**alakxender/dhivehi-vrd-images**](https://huggingface.co/datasets/alakxender/dhivehi-vrd-images) dataset.

* **Base model:** [unsloth/DeepSeek-OCR](https://huggingface.co/unsloth/DeepSeek-OCR)
* **More info on the model:** [deepseek-ai/DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR)
* **Dataset:** [alakxender/dhivehi-vrd-images](https://huggingface.co/datasets/alakxender/dhivehi-vrd-images)
* **Samples used:** 50k single-line Dhivehi sentences

## Usage

Inference using Huggingface transformers on NVIDIA GPUs. Requirements tested on python 3.12.9 + CUDA11.8：
```python
from transformers import AutoModel, AutoTokenizer
import torch
import os
os.environ["CUDA_VISIBLE_DEVICES"] = '0'
model_name = 'alakxender/deepseek-ocr-3b-vrd-dhivehi-50k-sl'

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, _attn_implementation='flash_attention_2', trust_remote_code=True, use_safetensors=True)
model = model.eval().cuda().to(torch.bfloat16)

# prompt = "<image>\nFree OCR. "
prompt = "<image>\nFree OCR. "
image_file = 'sl.png'
output_path = 'your/output/dir'

# infer(self, tokenizer, prompt='', image_file='', output_path = ' ', base_size = 1024, image_size = 640, crop_mode = True, test_compress = False, save_results = False):

# Tiny: base_size = 512, image_size = 512, crop_mode = False
# Small: base_size = 640, image_size = 640, crop_mode = False
# Base: base_size = 1024, image_size = 1024, crop_mode = False
# Large: base_size = 1280, image_size = 1280, crop_mode = False

# Gundam: base_size = 1024, image_size = 640, crop_mode = True

res = model.infer(tokenizer, prompt=prompt, image_file=image_file, output_path = output_path, base_size = 1024, image_size = 640, crop_mode=True, save_results = True, test_compress = True)
```