Instructions to use lkevinzc/Llama-3.2-3B-NuminaQA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lkevinzc/Llama-3.2-3B-NuminaQA with Transformers:

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

pipe = pipeline("text-generation", model="lkevinzc/Llama-3.2-3B-NuminaQA")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("lkevinzc/Llama-3.2-3B-NuminaQA")
model = AutoModelForCausalLM.from_pretrained("lkevinzc/Llama-3.2-3B-NuminaQA")

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use lkevinzc/Llama-3.2-3B-NuminaQA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lkevinzc/Llama-3.2-3B-NuminaQA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lkevinzc/Llama-3.2-3B-NuminaQA",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/lkevinzc/Llama-3.2-3B-NuminaQA

SGLang

How to use lkevinzc/Llama-3.2-3B-NuminaQA 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 "lkevinzc/Llama-3.2-3B-NuminaQA" \
    --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": "lkevinzc/Llama-3.2-3B-NuminaQA",
		"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 "lkevinzc/Llama-3.2-3B-NuminaQA" \
        --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": "lkevinzc/Llama-3.2-3B-NuminaQA",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use lkevinzc/Llama-3.2-3B-NuminaQA with Docker Model Runner:
```
docker model run hf.co/lkevinzc/Llama-3.2-3B-NuminaQA
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Llama-3.2-3B-NuminaQA

Introduction

This model serves as a 3B base in our minimalist R1-Zero recipe.

Training details:

Base model: HuggingFaceTB/FineMath-Llama-3B
Dataset: lkevinzc/numia-1.5-qa-concatenated
Epochs: 2
Learning rate: 1e-5

Citation

@article{liu2025understanding,
  title={Understanding r1-zero-like training: A critical perspective},
  author={Liu, Zichen and Chen, Changyu and Li, Wenjun and Qi, Penghui and Pang, Tianyu and Du, Chao and Lee, Wee Sun and Lin, Min},
  journal={arXiv preprint arXiv:2503.20783},
  year={2025}
}

Downloads last month: 20

Safetensors

Model size

3B params

Tensor type

F32

Model tree for lkevinzc/Llama-3.2-3B-NuminaQA

Base model

meta-llama/Llama-3.2-3B

Finetuned

HuggingFaceTB/FineMath-Llama-3B

Finetuned

(4)

this model

Finetunes

1 model

Quantizations

2 models

Dataset used to train lkevinzc/Llama-3.2-3B-NuminaQA

Paper for lkevinzc/Llama-3.2-3B-NuminaQA

Understanding R1-Zero-Like Training: A Critical Perspective

Paper • 2503.20783 • Published Mar 26, 2025 • 60

lkevinzc
/

Llama-3.2-3B-NuminaQA