Instructions to use viktoroo/SmolLM2-360M-Tools with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use viktoroo/SmolLM2-360M-Tools with Transformers:

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

pipe = pipeline("text-generation", model="viktoroo/SmolLM2-360M-Tools")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("viktoroo/SmolLM2-360M-Tools")
model = AutoModelForCausalLM.from_pretrained("viktoroo/SmolLM2-360M-Tools")
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]:]))

Transformers.js

How to use viktoroo/SmolLM2-360M-Tools with Transformers.js:

// npm i @huggingface/transformers
import { pipeline } from '@huggingface/transformers';

// Allocate pipeline
const pipe = await pipeline('text-generation', 'viktoroo/SmolLM2-360M-Tools');

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use viktoroo/SmolLM2-360M-Tools with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "viktoroo/SmolLM2-360M-Tools"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "viktoroo/SmolLM2-360M-Tools",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/viktoroo/SmolLM2-360M-Tools

SGLang

How to use viktoroo/SmolLM2-360M-Tools 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 "viktoroo/SmolLM2-360M-Tools" \
    --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": "viktoroo/SmolLM2-360M-Tools",
		"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 "viktoroo/SmolLM2-360M-Tools" \
        --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": "viktoroo/SmolLM2-360M-Tools",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use viktoroo/SmolLM2-360M-Tools with Docker Model Runner:
```
docker model run hf.co/viktoroo/SmolLM2-360M-Tools
```

viktoroo commited on Nov 2, 2025

Commit

a61e1bb

verified ·

1 Parent(s): 932aa54

Update README.md

Browse files

Files changed (1) hide show

README.md +14 -0

README.md CHANGED Viewed

@@ -17,6 +17,20 @@ base_model:
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/oWWfzW4RbWkVIo7f-5444.png)
 ##  Table of Contents
 1. [Model Summary](##model-summary)

 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/oWWfzW4RbWkVIo7f-5444.png)
+> [!WARNING]
+> # DISCLAIMER
+> This repository is a repack of `HuggingFaceTB/SmolLM2-360M-Instruct`.
+>
+> No model weights were modified. No additional fine-tuning was performed.
+> The only change is an updated `chat_template` in `tokenizer_config.json`.
+>
+> The new chat template:
+> - injects a system prompt if none is provided,
+> - documents available external tools,
+> - defines a function-calling protocol using `<tool_call>` and `<tool_response>` blocks.
+>
+> Because the weights are unchanged, behavior, capabilities, and safety limitations are inherited directly from the original `SmolLM2-360M-Instruct` checkpoint. The model has not been trained or aligned to actually follow the new tool-calling protocol. It may produce malformed tool calls or incorrect tool usage without further fine-tuning.
 ##  Table of Contents
 1. [Model Summary](##model-summary)