Instructions to use RabotniKuma/Fast-OpenMath-Nemotron-14B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RabotniKuma/Fast-OpenMath-Nemotron-14B with Transformers:

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

pipe = pipeline("text-generation", model="RabotniKuma/Fast-OpenMath-Nemotron-14B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RabotniKuma/Fast-OpenMath-Nemotron-14B")
model = AutoModelForCausalLM.from_pretrained("RabotniKuma/Fast-OpenMath-Nemotron-14B")
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

vLLM

How to use RabotniKuma/Fast-OpenMath-Nemotron-14B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RabotniKuma/Fast-OpenMath-Nemotron-14B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RabotniKuma/Fast-OpenMath-Nemotron-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RabotniKuma/Fast-OpenMath-Nemotron-14B

SGLang

How to use RabotniKuma/Fast-OpenMath-Nemotron-14B 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 "RabotniKuma/Fast-OpenMath-Nemotron-14B" \
    --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": "RabotniKuma/Fast-OpenMath-Nemotron-14B",
		"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 "RabotniKuma/Fast-OpenMath-Nemotron-14B" \
        --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": "RabotniKuma/Fast-OpenMath-Nemotron-14B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RabotniKuma/Fast-OpenMath-Nemotron-14B with Docker Model Runner:
```
docker model run hf.co/RabotniKuma/Fast-OpenMath-Nemotron-14B
```

Improve model card for Fast-OpenMath-Nemotron-14B

by nielsr HF Staff - opened Jul 16, 2025

base: refs/heads/main

←

from: refs/pr/2

Discussion Files changed

+97

-28

Improve model card for Fast-OpenMath-Nemotron-14Be79d352d

nielsr

Jul 16, 2025

Hello team,

I've opened this PR to significantly enhance the model card for Fast-OpenMath-Nemotron-14B.

The updates include:

Added the paper abstract for a more comprehensive overview of the research.
Expanded the model description by incorporating the detailed summary from the project's GitHub README, providing a clearer introduction.
Created a dedicated "Links" section to centralize access to the paper, GitHub repository, project page, and Kaggle discussion for easy navigation.
Included Hugging Face dataset links from the GitHub repository, making the training data easily discoverable and promoting reproducibility.
Added in-depth training details, covering the two-stage training process (SFT and GRPO), specific datasets used, and an explanation of the reward functions employed.
Included team attribution for proper recognition of the contributors.

These improvements aim to make the model card more informative, structured, and user-friendly, providing researchers with all necessary context and resources directly on the Hub.

Let me know if you have any questions or suggestions!
Thanks,
Niels

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment