Instructions to use moos124/sdft-modelv2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use moos124/sdft-modelv2 with Transformers:

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

pipe = pipeline("text-generation", model="moos124/sdft-modelv2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("moos124/sdft-modelv2")
model = AutoModelForCausalLM.from_pretrained("moos124/sdft-modelv2")
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 moos124/sdft-modelv2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "moos124/sdft-modelv2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "moos124/sdft-modelv2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/moos124/sdft-modelv2

SGLang

How to use moos124/sdft-modelv2 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 "moos124/sdft-modelv2" \
    --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": "moos124/sdft-modelv2",
		"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 "moos124/sdft-modelv2" \
        --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": "moos124/sdft-modelv2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use moos124/sdft-modelv2 with Docker Model Runner:
```
docker model run hf.co/moos124/sdft-modelv2
```

sdft-modelv2

File size: 2,029 Bytes

4e8d7f1

---
base_model: Qwen/Qwen2.5-1.5B-Instruct
library_name: transformers
model_name: sdft-modelv2
tags:
- generated_from_trainer
- sdft
- trl
- hf_jobs
licence: license
---

# Model Card for sdft-modelv2

This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct).
It has been trained using [TRL](https://github.com/huggingface/trl).

## Quick start

```python
from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="moos124/sdft-modelv2", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```

## Training procedure

 



This model was trained with SDFT, a method introduced in [Self-Training with On-Policy Self-Distillation for Language Model Alignment](https://huggingface.co/papers/2601.19897).

### Framework versions

- TRL: 1.3.0
- Transformers: 5.7.0
- Pytorch: 2.11.0
- Datasets: 4.8.5
- Tokenizers: 0.22.2

## Citations

Cite SDFT as:

```bibtex
@article{hubotter2026selftraining,
    title        = {{Self-Training with On-Policy Self-Distillation for Language Model Alignment}},
    author       = {Jonas H\"ubotter and Frederike L\"ubeck and Lejs Behric and Anton Baumann and Marco Bagatella and Daniel Marta and Ido Hakimi and Idan Shenfeld and Thomas Kleine Buening and Carlos Guestrin and Andreas Krause},
    year         = 2026,
    eprint       = {arXiv:2601.19897}
}
```

Cite TRL as:
    
```bibtex
@software{vonwerra2020trl,
  title   = {{TRL: Transformers Reinforcement Learning}},
  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
  license = {Apache-2.0},
  url     = {https://github.com/huggingface/trl},
  year    = {2020}
}
```