Instructions to use culturalheritagenus/Rumi-degarbler-Gemma-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use culturalheritagenus/Rumi-degarbler-Gemma-v1 with Transformers:

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

pipe = pipeline("text-generation", model="culturalheritagenus/Rumi-degarbler-Gemma-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("culturalheritagenus/Rumi-degarbler-Gemma-v1")
model = AutoModelForMultimodalLM.from_pretrained("culturalheritagenus/Rumi-degarbler-Gemma-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use culturalheritagenus/Rumi-degarbler-Gemma-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "culturalheritagenus/Rumi-degarbler-Gemma-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "culturalheritagenus/Rumi-degarbler-Gemma-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/culturalheritagenus/Rumi-degarbler-Gemma-v1

SGLang

How to use culturalheritagenus/Rumi-degarbler-Gemma-v1 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 "culturalheritagenus/Rumi-degarbler-Gemma-v1" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "culturalheritagenus/Rumi-degarbler-Gemma-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "culturalheritagenus/Rumi-degarbler-Gemma-v1" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "culturalheritagenus/Rumi-degarbler-Gemma-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use culturalheritagenus/Rumi-degarbler-Gemma-v1 with Docker Model Runner:
```
docker model run hf.co/culturalheritagenus/Rumi-degarbler-Gemma-v1
```

Model Card for Model ID

Model Details

Model Description

This model is trained to produce clean Jawi text from a garbled source, as part of an OCR pipeline.

Developed by: Computational Cultural Heritage Research Group (NUS).
Model type: Gemma 2 9B
Language(s) (NLP): Malay, English
Finetuned from model aisingapore/Gemma-SEA-LION-v3-9B-IT

How to Get Started with the Model

Use the code below to get started with the model:

from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer

trained_model = AutoModelForCausalLM.from_pretrained(
    "culturalheritagenus/rumi-degarbler-v1",
    device_map="auto",
    torch_dtype=torch.bfloat16
)
trained_tokenizer = AutoTokenizer.from_pretrained("culturalheritagenus/rumi-degarbler-v1")

To perform inference:

messages = [
    {"role": "user", "content": "You are a Malay language spelling corrector. I will give you some text written in messy Rumi (shortened or mistyped). Rewrite it in correct Malay Rumi spelling.\naurng ank. yngdim dimn anm aurngdan"},
]
inputs = tokenizer.apply_chat_template(
    messages,
    tokenize = True,
    add_generation_prompt = True, # Must add for generation
    return_tensors = "pt",
).to("cuda")


text_streamer = TextStreamer(tokenizer)
_ = trained_model.generate(input_ids = inputs, streamer = text_streamer, max_new_tokens = 128, use_cache = True)