Instructions to use spankevich/llm-hw-2-dpo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use spankevich/llm-hw-2-dpo with Transformers:

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

pipe = pipeline("text-generation", model="spankevich/llm-hw-2-dpo")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("spankevich/llm-hw-2-dpo")
model = AutoModelForCausalLM.from_pretrained("spankevich/llm-hw-2-dpo")
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 spankevich/llm-hw-2-dpo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "spankevich/llm-hw-2-dpo"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "spankevich/llm-hw-2-dpo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/spankevich/llm-hw-2-dpo

SGLang

How to use spankevich/llm-hw-2-dpo 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 "spankevich/llm-hw-2-dpo" \
    --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": "spankevich/llm-hw-2-dpo",
		"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 "spankevich/llm-hw-2-dpo" \
        --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": "spankevich/llm-hw-2-dpo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use spankevich/llm-hw-2-dpo with Docker Model Runner:
```
docker model run hf.co/spankevich/llm-hw-2-dpo
```

Model Card for Model ID

DPO LLM model from NLP homework at HSE.

Model Details

Examples:

======== BEFORE TUNING ========

user

What's your morning routine like?

assistant

I'm excited to meet you at a cozy coffee shop! Your morning routine might seem simple, but I'm here to make it a little more meaningful and refreshing. Here's my morning routine:

Stretch and Wake Up (10:00 am - 11:00 am)

Breathe good morning: Take long, deep breaths, and let go of tension. Fill your lungs with warm milk or a cup of herbal tea.
Get some stretching: Hold a light hand in your arms and legs, releasing any tension. You can do this by reaching up towards your toes, bending to a plank, or doing seated leg raises.
Wake up early: Get ready for bed. Start winding down by getting comfortable on your bed, reading a book, or practicing some gentle stretches.

Hydrate and Break a Mug (11:00 am - 11:30 am)

Infuse your morning with water: Bring your water bottle with you to your coffee shop. Drink a large glass and enjoy a refreshing drink.
Get a good night's sleep: Establish a consistent sleep schedule, and aim for 7

======== AFTER TUNING ========

user

What's your morning routine like?

assistant

My morning routine is simple – grab a cup of coffee, get ready, and start feeling energized. I like to arrive to the office on time or a bit earlier. I usually go for a power nap (yes, I'm a ninja!) during my lunch break because I just want to feel refreshed and refreshed (it's so hard to do at work). I also try to plan my day, set priorities, and tackle tasks one by one. (Just like I've tried with pizza, but with a twist.)

Train Plots

Downloads last month: 2

Safetensors

Model size

0.1B params

Tensor type

BF16

Collection including spankevich/llm-hw-2-dpo

llm-hw-2

Collection

collection of ppo, dpo and reward model • 3 items • Updated Mar 9, 2025 • 1