Instructions to use PracticalWork/SmolLM2-1.7B-Instruct-tuned with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PracticalWork/SmolLM2-1.7B-Instruct-tuned with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM2-1.7B-Instruct")
model = PeftModel.from_pretrained(base_model, "PracticalWork/SmolLM2-1.7B-Instruct-tuned")

Transformers

How to use PracticalWork/SmolLM2-1.7B-Instruct-tuned with Transformers:

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

pipe = pipeline("text-generation", model="PracticalWork/SmolLM2-1.7B-Instruct-tuned")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("PracticalWork/SmolLM2-1.7B-Instruct-tuned", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use PracticalWork/SmolLM2-1.7B-Instruct-tuned with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "PracticalWork/SmolLM2-1.7B-Instruct-tuned"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "PracticalWork/SmolLM2-1.7B-Instruct-tuned",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/PracticalWork/SmolLM2-1.7B-Instruct-tuned

SGLang

How to use PracticalWork/SmolLM2-1.7B-Instruct-tuned 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 "PracticalWork/SmolLM2-1.7B-Instruct-tuned" \
    --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": "PracticalWork/SmolLM2-1.7B-Instruct-tuned",
		"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 "PracticalWork/SmolLM2-1.7B-Instruct-tuned" \
        --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": "PracticalWork/SmolLM2-1.7B-Instruct-tuned",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use PracticalWork/SmolLM2-1.7B-Instruct-tuned with Docker Model Runner:
```
docker model run hf.co/PracticalWork/SmolLM2-1.7B-Instruct-tuned
```

SmolLM2-1.7B-Instruct-tuned

File size: 1,773 Bytes

---
library_name: peft
license: apache-2.0
base_model: HuggingFaceTB/SmolLM2-1.7B-Instruct
tags:
- base_model:adapter:HuggingFaceTB/SmolLM2-1.7B-Instruct
- transformers
pipeline_tag: text-generation
model-index:
- name: SmolLM2-1.7B-Instruct-tuned
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# SmolLM2-1.7B-Instruct-tuned

This model is a fine-tuned version of [HuggingFaceTB/SmolLM2-1.7B-Instruct](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4442
- Perplexity: 4.2384

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.003
- train_batch_size: 12
- eval_batch_size: 12
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Perplexity |
|:-------------:|:------:|:----:|:---------------:|:----------:|
| No log        | 0      | 0    | 3.1117          | 22.4598    |
| No log        | 0.6011 | 333  | 1.5821          | 4.8650     |
| 1.6345        | 1.2022 | 666  | 1.5242          | 4.5915     |
| 1.6345        | 1.8032 | 999  | 1.4899          | 4.4364     |
| 1.4894        | 2.4043 | 1332 | 1.4634          | 4.3205     |
| 1.4894        | 3      | 1664 | 1.4442          | 4.2384     |


### Framework versions

- PEFT 0.16.0
- Transformers 4.54.1
- Pytorch 2.7.1+cu128
- Datasets 4.0.0
- Tokenizers 0.21.4