Ouro-1.4B-Thinking

---

Fork Description

• Makes the Ouro model compatible with Transformers v5 (tested with 5.9.0).
• Fixes several smaller bugs.
• Serve Ouro with adaptive exit efficiently with Loop-SGLang.

---

Ouro-1.4B-Thinking

📚 Paper • 🏠 Project Page

Model Description

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

Ouro-1.4B-Thinking is a reasoning-specialized variant of the Ouro-1.4B base model, enhanced through supervised fine-tuning on high-quality reasoning data.

Key Features

• Advanced Reasoning: Specifically optimized for mathematical and scientific reasoning tasks
• Compact Size: Competitive with 4B models despite having only 1.4B 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:

{
  "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. When using the Transformers library, the model will always execute the full number of total_ut_steps and select the exited loop state afterwards. Later tokens can therefore attend to KV entries from recurrent depths that earlier tokens would not have reached under actual early exit. Use Loop-SGLang for last-exited KV and early-exit compute savings.

Example: Modify recurrent steps

from transformers import AutoConfig, AutoModelForCausalLM

config = AutoConfig.from_pretrained("KristianS7/Ouro-1.4B-Thinking", trust_remote_code=True)
config.total_ut_steps = 3  # Use 3 recurrent steps instead of 4
model = AutoModelForCausalLM.from_pretrained(
    "KristianS7/Ouro-1.4B-Thinking",
    config=config,
    device_map="auto",
    trust_remote_code=True
)

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. For adaptive exit use, Loop-SGLang instead.

Model Architecture

Based on Ouro-1.4B with additional reasoning fine-tuning:

Training Details

Pre-training

• Training Tokens: 7.7T tokens across 4 stages
• Base Architecture: Ouro-1.4B

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: Works with transformers==5.9.0.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "KristianS7/Ouro-1.4B-Thinking"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    device_map="auto",
    torch_dtype="auto",
    trust_remote_code=True
)

# 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",
    enable_thinking=True
).to(model.device)

outputs = model.generate(inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Acknowledgments

We thank @Antizana for the KV cache fix merged from ouro-cache-fix, which resolved a critical compatibility issue with transformers>=4.56.0.

Citation

@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.