Instructions to use Serialtechlab/dhivehi-trocr-small-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Serialtechlab/dhivehi-trocr-small-v2 with Transformers:

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

pipe = pipeline("image-text-to-text", model="Serialtechlab/dhivehi-trocr-small-v2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForImageTextToText

tokenizer = AutoTokenizer.from_pretrained("Serialtechlab/dhivehi-trocr-small-v2")
model = AutoModelForImageTextToText.from_pretrained("Serialtechlab/dhivehi-trocr-small-v2")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Serialtechlab/dhivehi-trocr-small-v2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Serialtechlab/dhivehi-trocr-small-v2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Serialtechlab/dhivehi-trocr-small-v2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Serialtechlab/dhivehi-trocr-small-v2

SGLang

How to use Serialtechlab/dhivehi-trocr-small-v2 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 "Serialtechlab/dhivehi-trocr-small-v2" \
    --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": "Serialtechlab/dhivehi-trocr-small-v2",
		"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 "Serialtechlab/dhivehi-trocr-small-v2" \
        --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": "Serialtechlab/dhivehi-trocr-small-v2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Serialtechlab/dhivehi-trocr-small-v2 with Docker Model Runner:
```
docker model run hf.co/Serialtechlab/dhivehi-trocr-small-v2
```

You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

dhivehi-trocr-small-v2

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0533
Cer: 0.1894

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.1992	0.2350	5000	0.2111	0.1751
0.1335	0.4700	10000	0.1193	0.1761
0.1055	0.7050	15000	0.0941	0.1838
0.0928	0.9400	20000	0.0898	0.1903
0.0806	1.1750	25000	0.0911	0.1823
0.0698	1.4100	30000	0.0699	0.1954
0.0666	1.6450	35000	0.0662	0.1899
0.0619	1.8801	40000	0.0674	0.1960
0.0577	2.1151	45000	0.0584	0.1895
0.0579	2.3501	50000	0.0686	0.1932
0.0515	2.5851	55000	0.0551	0.1892
0.0528	2.8201	60000	0.0533	0.1894

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 16

Safetensors

Model size

28.9M params

Tensor type

F32

Inference Providers NEW

Image-Text-to-Text

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