Instructions to use scpalmetto/Ouro-2.6B-Thinking-Fixed with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with Transformers:

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

pipe = pipeline("text-generation", model="scpalmetto/Ouro-2.6B-Thinking-Fixed", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("scpalmetto/Ouro-2.6B-Thinking-Fixed", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("scpalmetto/Ouro-2.6B-Thinking-Fixed", trust_remote_code=True, 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

llama.cpp

How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

Use Docker

docker model run hf.co/scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

LM Studio
Jan

vLLM

How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "scpalmetto/Ouro-2.6B-Thinking-Fixed"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "scpalmetto/Ouro-2.6B-Thinking-Fixed",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

SGLang

How to use scpalmetto/Ouro-2.6B-Thinking-Fixed 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 "scpalmetto/Ouro-2.6B-Thinking-Fixed" \
    --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": "scpalmetto/Ouro-2.6B-Thinking-Fixed",
		"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 "scpalmetto/Ouro-2.6B-Thinking-Fixed" \
        --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": "scpalmetto/Ouro-2.6B-Thinking-Fixed",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with Ollama:
```
ollama run hf.co/scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with Docker Model Runner:
```
docker model run hf.co/scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M
```

Lemonade

How to use scpalmetto/Ouro-2.6B-Thinking-Fixed with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull scpalmetto/Ouro-2.6B-Thinking-Fixed:Q4_K_M

Run and chat with the model

lemonade run user.Ouro-2.6B-Thinking-Fixed-Q4_K_M

List all available models

lemonade list

Atomic Chat

Ouro-2.6B-Thinking-Fixed / README.md

scpalmetto

Upload README.md with huggingface_hub

6a6b306 verified 7 months ago

preview code

Raw History Blame Contribute Delete

5.01 kB

	---
	library_name: transformers
	license: apache-2.0
	pipeline_tag: text-generation
	tags:
	- looped-language-model
	- reasoning
	- recurrent-depth
	- thinking
	- chain-of-thought
	---

	# Ouro-2.6B-Thinking

	![Ouro Logo](assets/logo.png)

	## Model Description


	⚠️ IMPORTANT: This model is intended for research purposes only. It is provided as-is without warranties for production use.


	Ouro-2.6B-Thinking is a reasoning-specialized variant of the Ouro-2.6B base model, enhanced through supervised fine-tuning on high-quality reasoning data. Please use ``transformers==4.54.1``for compatibility.

	![Thinking Model Performance](assets/ouro_thinking.png)

	## Key Features

	- Advanced Reasoning: Specifically optimized for mathematical and scientific reasoning tasks
	- Compact Size: Competitive with 4B models despite having only 2.6B parameters
	- Cross-Step Consistency: Intermediate recurrent outputs can serve as reliable proxies for final answers
	- Explicit Thinking Process: Trained to generate detailed reasoning steps

	## Configuration

	### Recurrent Steps and Adaptive Exit

	The model's computational behavior can be configured through the `config.json` file:

	```json
	{
	"total_ut_steps": 4,
	"early_exit_threshold": 1.0
	}
	```

	- `total_ut_steps`: Controls the number of recurrent steps (default: 4). You can adjust this value to trade off between performance and computation time.
	- `early_exit_threshold`: Controls the adaptive exit mechanism (default: 1.0). Lower values encourage earlier exit, while 1.0 means always use all steps.

	Example: Modify recurrent steps
	```python
	from transformers import AutoConfig, AutoModelForCausalLM

	config = AutoConfig.from_pretrained("ByteDance/Ouro-2.6B-Thinking")
	config.total_ut_steps = 3 # Use 3 recurrent steps instead of 4
	model = AutoModelForCausalLM.from_pretrained(
	"ByteDance/Ouro-2.6B-Thinking",
	config=config,
	device_map="auto"
	)
	```

	> Note: vLLM does not currently support the adaptive exit feature due to its inference optimization characteristics. When using vLLM, the model will always execute the full number of `total_ut_steps`.


	## Model Architecture

	Based on Ouro-2.6B with additional reasoning fine-tuning:

	\| Configuration \| Value \|
	\|:---\|:---\|
	\| Parameters \| 2.6B \|
	\| Layers \| 24 \|
	\| Recurrent Steps \| 4 \|
	\| Hidden Size \| 2048 \|
	\| Attention Heads \| Multi-Head Attention (MHA) \|
	\| FFN Activation \| SwiGLU \|
	\| Position Embedding \| RoPE \|
	\| Vocabulary Size \| 49,152 \|
	\| Context Length \| 32K (SFT) \|
	\| Normalization \| Sandwich RMSNorm \|

	## Training Details

	### Pre-training
	- Training Tokens: 7.7T tokens across 4 stages
	- Base Architecture: Ouro-2.6B

	### Supervised Fine-Tuning
	- Data Size: ~8.3M examples
	- Data Composition:
	- Mathematics: 3.5M examples (OpenThoughts3, AceReason-1.1-SFT)
	- Code: 3.2M examples (AceReason, OpenCodeReasoning, Llama-Nemotron, OpenThoughts3)
	- Science: 808K examples (OpenThoughts3, Llama-Nemotron)
	- Chat: 767K examples (DeepWriting-20K)
	- Training: 2 epochs, max sequence length 32K
	- Optimizer: Adam (lr=2×10⁻⁵, β=(0.9, 0.95))
	- Scheduler: Cosine decay


	## Quick Start

	⚠️ IMPORTANT: Please use `transformers<4.56.0` to avoid compatibility issues. We recommend `transformers==4.54.1` or earlier versions.

	```python
	from transformers import AutoModelForCausalLM, AutoTokenizer

	model_name = "Bytedance/Ouro-2.6B-Thinking"
	tokenizer = AutoTokenizer.from_pretrained(model_name)
	model = AutoModelForCausalLM.from_pretrained(
	model_name,
	device_map="auto",
	torch_dtype="auto"
	)

	# Generate with reasoning
	messages = [
	{"role": "user", "content": "Solve: If 2x + 3 = 11, what is x?"}
	]
	inputs = tokenizer.apply_chat_template(
	messages,
	tokenize=True,
	add_generation_prompt=True,
	return_tensors="pt"
	).to(model.device)

	outputs = model.generate(inputs, max_new_tokens=512, temperature=1.0, top_p=0.7)
	print(tokenizer.decode(outputs[0], skip_special_tokens=True))
	```



	## Acknowledgments

	We thank [@Antizana](https://github.com/Antizana) for the KV cache fix merged from [ouro-cache-fix](https://github.com/Antizana/ouro-cache-fix), which resolved a critical compatibility issue with transformers>=4.56.0.

	## Citation

	```bibtex
	@article{zhu2025scaling,
	title={Scaling Latent Reasoning via Looped Language Models},
	author={Zhu, Rui-Jie and Wang, Zixuan and Hua, Kai and Zhang, Tianyu and Li, Ziniu and Que, Haoran and Wei, Boyi and Wen, Zixin and Yin, Fan and Xing, He and others},
	journal={arXiv preprint arXiv:2510.25741},
	year={2025}
	}


	## License

	This model is licensed under Apache-2.0. See the LICENSE file for details.

	## Project Links

	- Paper: [Scaling Latent Reasoning via Looped Language Models](https://huggingface.co/papers/2510.25741)
	- Project Page: [https://ouro-llm.github.io](https://ouro-llm.github.io)
	- Code: [https://github.com/ByteDance/Ouro](https://github.com/ByteDance/Ouro)

	---