Instructions to use oopere/martra-open-gemma-2b-it-dpo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use oopere/martra-open-gemma-2b-it-dpo with Transformers:

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

pipe = pipeline("text-generation", model="oopere/martra-open-gemma-2b-it-dpo")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("oopere/martra-open-gemma-2b-it-dpo")
model = AutoModelForCausalLM.from_pretrained("oopere/martra-open-gemma-2b-it-dpo", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use oopere/martra-open-gemma-2b-it-dpo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "oopere/martra-open-gemma-2b-it-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": "oopere/martra-open-gemma-2b-it-dpo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/oopere/martra-open-gemma-2b-it-dpo

SGLang

How to use oopere/martra-open-gemma-2b-it-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 "oopere/martra-open-gemma-2b-it-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": "oopere/martra-open-gemma-2b-it-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 "oopere/martra-open-gemma-2b-it-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": "oopere/martra-open-gemma-2b-it-dpo",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use oopere/martra-open-gemma-2b-it-dpo with Docker Model Runner:
```
docker model run hf.co/oopere/martra-open-gemma-2b-it-dpo
```

martra-open-gemma-2b-it-dpo / README.md

oopere

Upload GemmaForCausalLM

ec47410 verified about 2 years ago

preview code

Raw

History Blame Contribute Delete

1.98 kB

	---
	language:
	- en
	license: mit
	library_name: transformers
	tags:
	- dpo
	- 2b
	- gemma
	datasets:
	- argilla/distilabel-capybara-dpo-7k-binarized
	pipeline_tag: text-generation
	widget:
	- text: "3713841893836/4? \nLimit your response to mathematical expressions and symbols."
	example_title: 'Return only numbers. '
	- text: A group of 10 people is split into 3 different committees of 3, 4, and 3 people,
	respectively. In how many ways can this be done?
	example_title: Solve Problem
	---

	You can see the process with instructions for creating the model in the notebook: [Aligning_DPO_gemma-2b.ipynb](https://github.com/peremartra/Large-Language-Model-Notebooks-Course/blob/main/P2-MHF/Aligning_DPO_open_gemma-2b-it.ipynb)

	To create it, we started with the [google/gemma-2b-it](https://huggingface.co/google/gemma-2b-it) model and applied DPO alignment using the [distilabel-capybara-dpo-7k-binarized dataset](https://huggingface.co/datasets/argilla/distilabel-capybara-dpo-7k-binarized).

	Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.

	Row in the Dataset:

	*{'prompt': '<\|user\|>\nAssist me in calculating 9319357631 plus 595. Numbers and symbols only, please.<\|end\|>\n<\|assistant\|>\n',
	'chosen': 'The sum of 9319357631 and 595 is 9319358226.<\|end\|>\n',
	'rejected': 'The result of adding 9319357631 and 595 is 9319363626.<\|end\|>\n'}*

	Prompt:
	*3713841893836/4?
	Limit your response to mathematical expressions and symbols.*

	Response from the Base model:

	*To find the result of the division, we can simply divide the given number by 4:
	$$
	\frac{3713841893836}{4} = 928460473459*

	Response from the fine-tuned model:

	3713841893836 ÷ 4 = 928460473459


	If you want to see how the model was created, you can check out the [repository](https://github.com/peremartra/Large-Language-Model-Notebooks-Course) where the book's notebooks are kept up-to-date.