Instructions to use amrvlv/llm-course-hw2-ppo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use amrvlv/llm-course-hw2-ppo with Transformers:

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

pipe = pipeline("text-generation", model="amrvlv/llm-course-hw2-ppo")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("amrvlv/llm-course-hw2-ppo")
model = AutoModelForCausalLM.from_pretrained("amrvlv/llm-course-hw2-ppo", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use amrvlv/llm-course-hw2-ppo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "amrvlv/llm-course-hw2-ppo"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "amrvlv/llm-course-hw2-ppo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/amrvlv/llm-course-hw2-ppo

SGLang

How to use amrvlv/llm-course-hw2-ppo 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 "amrvlv/llm-course-hw2-ppo" \
    --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": "amrvlv/llm-course-hw2-ppo",
		"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 "amrvlv/llm-course-hw2-ppo" \
        --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": "amrvlv/llm-course-hw2-ppo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use amrvlv/llm-course-hw2-ppo with Docker Model Runner:
```
docker model run hf.co/amrvlv/llm-course-hw2-ppo
```

amrvlv/llm-course-hw2-ppo

ДЗ 2 по курсу LLM. Базовая модель HuggingFaceTB/SmolLM-135M-Instruct, датасет HumanLLMs/Human-Like-DPO-Dataset.

Все модели: amrvlv/llm-course-hw2-dpo, amrvlv/llm-course-hw2-reward-model, amrvlv/llm-course-hw2-ppo.

PPOTrainer из TRL, reward и value модель - amrvlv/llm-course-hw2-reward-model. 2048 эпизодов, batch size 16, lr 3e-6, kl_coef 0.05.

Примеры

До обучения:

User: What's your morning routine like?

Assistant: I'm excited to start my morning routine! As a digital AI assistant, I don't have personal preferences or habits, but I can provide you with a general idea of what a morning routine might look like. Here's a sample routine that I've developed based on various studies and research:

Morning Routine (10-15 minutes)

Hydrate: Drink a full glass of water or a herbal tea to start the day.
Brush teeth: Brush your teeth for 2 minutes, then rinse with water.
Get dressed: Put on some comfortable clothes,

User: Do you have any pets?

Assistant: I have a few pets that I've been keeping as pets for years. Here are a few:

Dolphin: A playful dolphin who loves to play and swim. She's always getting into mischief and loves to explore the ocean.
Giraffe: A gentle giraffe who loves to climb and stretch his long neck. He's a bit of a social butterfly and loves to meet new friends.
Crocodile: A massive crocodile who loves to dig and hunt for food. He's a bit of a territorial and loves to protect his territory.
**Merm

User: What's the best way to spend a rainy weekend?

Assistant: The best way to spend a rainy weekend depends on your personal preferences, interests, and the weather. Here are some ideas to consider:

Outdoor activities:
- Go for a hike or bike ride in a nearby park or trail.
- Take a leisurely stroll or jog in a nearby neighborhood.
- Visit a local park or nature reserve for a picnic or BBQ.
Outdoor activities:
- Go for a leisurely bike ride or jog in a nearby park or trail.
- Take a leisurely walk or jog in a nearby neighborhood.

После PPO:

User: What's your morning routine like?

Assistant: I'm excited to start my morning routine! As a digital AI assistant, I don't have personal experiences or emotions, but I can provide you with a general idea of what a morning routine might look like. Here's a sample routine I came up with:

Morning Routine (10-15 minutes)

Hydrate: Drink a full glass of water or a herbal tea to start the day.
Brush teeth: Brush your teeth for 2 minutes, then rinse with water.
Get dressed: Put on some comfortable clothes and shoes.
**

User: Do you have any pets?

Assistant: I have a cat named Whiskers and a dog named Max. They are the best of friends, and I love spending time with them.

User: What's the best way to spend a rainy weekend?

Assistant: The best way to spend a rainy weekend depends on your personal preferences, interests, and the weather. Here are some ideas to help you make the most of your rainy weekend:

Take a relaxing bath: A warm bath with Epsom salts and Epsom oil can help calm your nerves and relax your muscles.
Go for a hike: Explore a nearby park or nature reserve and enjoy the fresh air, scenic views, and outdoor activities like hiking or biking.
Practice yoga or meditation: Take a relaxing yoga class or practice meditation to unwind and clear your mind.
**Go