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

Libraries

How to use nvidia/OpenMath-Nemotron-14B-Kaggle with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("nvidia/OpenMath-Nemotron-14B-Kaggle")
model = AutoModelForCausalLM.from_pretrained("nvidia/OpenMath-Nemotron-14B-Kaggle")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nvidia/OpenMath-Nemotron-14B-Kaggle with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/nvidia/OpenMath-Nemotron-14B-Kaggle

SGLang

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

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

igitman commited on Apr 30, 2025

Commit

a1ddb28

verified ·

1 Parent(s): 1f21285

Update README.md

Browse files

Files changed (1) hide show

README.md +7 -39

README.md CHANGED Viewed

@@ -61,45 +61,13 @@ The pipeline we used to produce the data and models is fully open-sourced!
 We provide [all instructions](https://nvidia.github.io/NeMo-Skills/openmathreasoning1/)
 to fully reproduce our results, including data generation.
-# How to use the models?
-Our models can be used in 3 inference modes: chain-of-thought (CoT), tool-integrated reasoning (TIR) and generative solution selection (GenSelect).
-To run inference with CoT mode, you can use this example code snippet.
-```python
-import transformers
-import torch
-model_id = "nvidia/OpenMath-Nemotron-14B-Kaggle"
-pipeline = transformers.pipeline(
-    "text-generation",
-    model=model_id,
-    model_kwargs={"torch_dtype": torch.bfloat16},
-    device_map="auto",
-)
-messages = [
-    {
-        "role": "user",
-        "content": "Solve the following math problem. Make sure to put the answer (and only answer) inside \\boxed{}.\n\n" +
-        "What is the minimum value of $a^2+6a-7$?"},
-]
-outputs = pipeline(
-    messages,
-    max_new_tokens=4096,
-)
-print(outputs[0]["generated_text"][-1]['content'])
-```
-To run inference with TIR or GenSelect modes, we highly recommend to use our
 [reference implementation in NeMo-Skills](https://nvidia.github.io/NeMo-Skills/openmathreasoning1/evaluation/).
 Please note that these models have not been instruction tuned on general data and thus might not provide good answers outside of math domain.
 ## Citation
 If you find our work useful, please consider citing us!
@@ -134,7 +102,7 @@ This model is intended to facilitate research in the area of mathematical reason
 Huggingface 04/23/2025 <br>
-## Model Architecture: <br>
 **Architecture Type:** Transformer decoder-only language model  <br>
@@ -145,7 +113,7 @@ Huggingface 04/23/2025 <br>
 ** This model has 1.5B of model parameters. <br>
-## Input: <br>
 **Input Type(s):** Text <br>
@@ -157,7 +125,7 @@ Huggingface 04/23/2025 <br>
-## Output: <br>
 **Output Type(s):** Text <br>
@@ -173,7 +141,7 @@ Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated sys
-## Software Integration : <br>
 **Runtime Engine(s):** <br>
@@ -195,7 +163,7 @@ Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated sys
-## Model Version(s):
 [OpenMath-Nemotron-1.5B](https://huggingface.co/nvidia/OpenMath-Nemotron-1.5B)

 We provide [all instructions](https://nvidia.github.io/NeMo-Skills/openmathreasoning1/)
 to fully reproduce our results, including data generation.
+## How to use the models?
+This model will always use code execution to solve math problems, so we highly recommend to run inference with our
 [reference implementation in NeMo-Skills](https://nvidia.github.io/NeMo-Skills/openmathreasoning1/evaluation/).
 Please note that these models have not been instruction tuned on general data and thus might not provide good answers outside of math domain.
 ## Citation
 If you find our work useful, please consider citing us!
 Huggingface 04/23/2025 <br>
+### Model Architecture: <br>
 **Architecture Type:** Transformer decoder-only language model  <br>
 ** This model has 1.5B of model parameters. <br>
+### Input: <br>
 **Input Type(s):** Text <br>
+### Output: <br>
 **Output Type(s):** Text <br>
+### Software Integration : <br>
 **Runtime Engine(s):** <br>
+### Model Version(s):
 [OpenMath-Nemotron-1.5B](https://huggingface.co/nvidia/OpenMath-Nemotron-1.5B)