Instructions to use SynastriaNetworks/Uirapuru-U1-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SynastriaNetworks/Uirapuru-U1-9B with Transformers:

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

pipe = pipeline("image-text-to-text", model="SynastriaNetworks/Uirapuru-U1-9B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("SynastriaNetworks/Uirapuru-U1-9B")
model = AutoModelForMultimodalLM.from_pretrained("SynastriaNetworks/Uirapuru-U1-9B", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use SynastriaNetworks/Uirapuru-U1-9B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SynastriaNetworks/Uirapuru-U1-9B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SynastriaNetworks/Uirapuru-U1-9B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/SynastriaNetworks/Uirapuru-U1-9B

SGLang

How to use SynastriaNetworks/Uirapuru-U1-9B 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 "SynastriaNetworks/Uirapuru-U1-9B" \
    --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": "SynastriaNetworks/Uirapuru-U1-9B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "SynastriaNetworks/Uirapuru-U1-9B" \
        --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": "SynastriaNetworks/Uirapuru-U1-9B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use SynastriaNetworks/Uirapuru-U1-9B with Docker Model Runner:
```
docker model run hf.co/SynastriaNetworks/Uirapuru-U1-9B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Uirapuru U1

A post-trained version of Qwen3.5-9B-Base focused on improving instruction following, Brazilian Portuguese, English, reasoning, and practical assistant capabilities.

Overview

Uirapuru U1 is a post-trained large language model developed by SynastrIA Networks. Instead of training a model from scratch, U1 builds upon the strong foundation of Qwen3.5-9B-Base and specializes it through supervised post-training.

The project prioritizes:

High-quality instruction following
Natural Brazilian Portuguese
Strong English performance
Better reasoning
Tool calling
Helpful, safe and practical conversations

Base Model

Base: Qwen3.5-9B-Base
Architecture: Decoder-only Transformer
Training Type: Post-training (Supervised Fine-Tuning)

Languages

Primary support:

Portuguese (Brazil)
English

General multilingual capabilities are inherited from the base model.

Training

Uirapuru U1 was post-trained using a curated instruction dataset composed of Portuguese and English conversations.

Training objectives include:

Instruction following
Question answering
Reasoning
Code understanding
Tool calling
Conversation quality
Knowledge formatting
Writing assistance

Intended Uses

Uirapuru U1 is intended for:

AI assistants
Chatbots
Coding assistance
Educational applications
Content generation
Summarization
Translation
Tool-calling agents
General NLP research

Limitations

Like any language model, Uirapuru U1 may:

Produce incorrect or outdated information.
Make reasoning mistakes.
Reflect biases inherited from training data.
Generate confident but inaccurate responses.

Human verification is recommended for important or high-risk tasks.

Evaluation

During development, emphasis was placed on:

Instruction adherence
Portuguese fluency
English quality
Reasoning performance
Response consistency
Tool calling reliability

Formal benchmark results will be published in future releases.

Safety

Uirapuru U1 was designed to be a helpful assistant while avoiding harmful or unsafe outputs whenever possible.

The model should not be relied upon as the sole source for legal, financial or medical decisions.

Citation

@misc{uirapuruu1,
  title={Uirapuru U1},
  author={SynastrIA Networks},
  year={2026},
  howpublished={Hugging Face}
}

Acknowledgements

Uirapuru U1 is built upon the excellent Qwen3.5-9B-Base model developed by the Qwen team.

Special thanks to the open-source AI community for the tools, libraries and research that made this project possible.

License

This repository follows the license and usage requirements of the base model. Please review the original Qwen license before deployment.

Downloads last month: 33

Safetensors

Model size

10B params

Tensor type

BF16

F32

Model tree for SynastriaNetworks/Uirapuru-U1-9B

Base model

Qwen/Qwen3.5-9B-Base

Finetuned

(144)

this model

Quantizations

2 models

SynastriaNetworks
/

Uirapuru-U1-9B