Instructions to use theprint/Coma-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use theprint/Coma-3B with Transformers:

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

pipe = pipeline("text-generation", model="theprint/Coma-3B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("theprint/Coma-3B")
model = AutoModelForCausalLM.from_pretrained("theprint/Coma-3B", 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 theprint/Coma-3B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "theprint/Coma-3B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "theprint/Coma-3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/theprint/Coma-3B

SGLang

How to use theprint/Coma-3B 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 "theprint/Coma-3B" \
    --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": "theprint/Coma-3B",
		"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 "theprint/Coma-3B" \
        --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": "theprint/Coma-3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio

How to use theprint/Coma-3B with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for theprint/Coma-3B to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for theprint/Coma-3B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for theprint/Coma-3B to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="theprint/Coma-3B",
    max_seq_length=2048,
)

Docker Model Runner
How to use theprint/Coma-3B with Docker Model Runner:
```
docker model run hf.co/theprint/Coma-3B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Coma 3B

Coma is based on Qwen 2.5 3B, GRPO-fine tuned on the natural reasoning data set from Meta.

GGUF

There are quantized versions available at theprint/Coma-3B-GGUF in GGUF format.

Testing

The following system prompt was used in testing of the model:

Between the tags <think> and </think>: You will first think through your answer carefully step by step, including any potential risks or pit falls, the context of the user's request, and how best to present this. These are notes for yourself, so be detailed and honest in your assessment.

When you are done with the analysis, you must use your own guidelines from the previous section to construct the final response. The user will only see this section.

In summary, respond using this pattern:
<think>
  First, ...
</think>
  [your response]

Testing was done at temperature=1.0, top_k=45 and top_p=0.95.

Developed by: theprint
License: apache-2.0
Finetuned from model : unsloth/qwen2.5-3b-instruct-unsloth-bnb-4bit

This qwen2 model was trained 2x faster with Unsloth and Huggingface's TRL library.

Downloads last month: 8

Safetensors

Model size

3B params

Tensor type

BF16

Model tree for theprint/Coma-3B

Quantizations

3 models

theprint
/

Coma-3B

Coma 3B

GGUF

Testing

Model tree for theprint/Coma-3B

Dataset used to train theprint/Coma-3B