Instructions to use mevsg/rumi-correction-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mevsg/rumi-correction-v1 with Transformers:

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

pipe = pipeline("text-generation", model="mevsg/rumi-correction-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mevsg/rumi-correction-v1")
model = AutoModelForCausalLM.from_pretrained("mevsg/rumi-correction-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 mevsg/rumi-correction-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mevsg/rumi-correction-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": "mevsg/rumi-correction-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mevsg/rumi-correction-v1

SGLang

How to use mevsg/rumi-correction-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 "mevsg/rumi-correction-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": "mevsg/rumi-correction-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 "mevsg/rumi-correction-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": "mevsg/rumi-correction-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mevsg/rumi-correction-v1 with Docker Model Runner:
```
docker model run hf.co/mevsg/rumi-correction-v1
```

Model Card for Model ID

Model Details

Model Description

This model is trained with QLoRA with parameters r = lora_alpha = 4.

Developed by: hyhyhyhyyhyh
Model type: Gemma 2 9B
Language(s) (NLP): Malay, English
License: [More Information Needed]
Finetuned from model aisingapore/Gemma-SEA-LION-v3-9B-IT

Model Sources [optional]

Repository: [More Information Needed]
Paper [optional]: [More Information Needed]
Demo [optional]: [More Information Needed]

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-correction-v1.1",
    device_map="auto",
    torch_dtype=torch.bfloat16
)
trained_tokenizer = AutoTokenizer.from_pretrained("culturalheritagenus/rumi-correction-v1.1")

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)

Training Details

Training Data

The model was trained on culturalheritagenus/rumi-correction-v1.1-data-v3

Training Procedure

To replicate this model, please refer to the provided script and below. Ensure that the versions of all languages and libraries are the same.

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

Hardware Type: 1x GH200 (96 GB)
Hours used: ~12
Cloud Provider: Lambda
Compute Region: US-East (Lambda Labs)

Technical Specifications

Software

Python version: 3.10.12
CUDA version: 12.8
Torch version: 2.7.1+cu128

Citation [optional]

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Model Card Authors [optional]

hyhyhyhyyhyh

Downloads last month: 5

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for mevsg/rumi-correction-v1

Base model

google/gemma-2-9b

Finetuned

aisingapore/Gemma-SEA-LION-v3-9B

Finetuned

aisingapore/Gemma-SEA-LION-v3-9B-IT

Finetuned

(3)

this model

Paper for mevsg/rumi-correction-v1

Quantifying the Carbon Emissions of Machine Learning

Paper • 1910.09700 • Published Oct 21, 2019 • 50