> For the complete documentation index, see [llms.txt](https://hibernates-ai.gitbook.io/hibernates-2b-r1-v1/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hibernates-ai.gitbook.io/hibernates-2b-r1-v1/hibernates-hibernates-2b-r1-v1.md).

# Hibernates/Hibernates-2B-R1-V1

A highly efficient 2B parameter language model optimized for reasoning and dialogue tasks.

### Model Overview <a href="#model-overview" id="model-overview"></a>

Hibernates-2B is a custom transformer architecture designed for advanced language understanding and generation. Built with performance and efficiency in mind, it leverages state-of-the-art techniques for natural language processing.

#### Key Features <a href="#key-features" id="key-features"></a>

* 2B Parameters
* 4096 Token Context Window
* Custom Transformer Architecture
* Optimized for CPU and GPU Inference
* Multi-Turn Dialogue Support

### Technical Specifications <a href="#technical-specifications" id="technical-specifications"></a>

* **Architecture**: Custom Transformer
* **Parameters**: 2 Billion
* **Context Length**: 4096 tokens
* **Model Type**: Decoder-only
* **Tokenizer**: Custom WordPiece
* **Format**: SafeTensors

### Usage Guide <a href="#usage-guide" id="usage-guide"></a>

```
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load model and tokenizer
model_id = "Hibernates-2B-R1-V1"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto"
)

# Example conversation
messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "How can you help me today?"}
]

# Generate response
input_text = tokenizer.apply_chat_template(messages, tokenize=False)
inputs = tokenizer(input_text, return_tensors="pt").to(model.device)
outputs = model.generate(
    inputs["input_ids"],
    max_new_tokens=512,
    temperature=0.7,
    top_p=0.95
)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
```

### Performance Characteristics <a href="#performance-characteristics" id="performance-characteristics"></a>

#### Strengths <a href="#strengths" id="strengths"></a>

* Efficient Resource Usage
* Strong Reasoning Capabilities
* Multi-Turn Dialogue
* Context Awareness
* Instruction Following

#### Considerations <a href="#considerations" id="considerations"></a>

* Resource Requirements: 8GB+ GPU RAM recommended
* Task Specificity: Best suited for dialogue and reasoning tasks
* Language Support: Primary focus on English
* Model Size: Optimized for balance of performance and efficiency

### License and Usage <a href="#license-and-usage" id="license-and-usage"></a>

* Research and commercial use permitted
* Attribution appreciated but not required
* No warranty provided

### Citation <a href="#citation" id="citation"></a>

If you use this model in your research, please cite:

```
@software{hibernates2b_2024,
  title={Hibernates-2B: Efficient Language Model for Reasoning},
  year={2024},
  version={R1-V1}
}
```

### Acknowledgments <a href="#acknowledgments" id="acknowledgments"></a>

Built using PyTorch and Hugging Face Transformers. Special thanks to the open-source AI community.

### Download Instructions <a href="#download-instructions" id="download-instructions"></a>

Due to file size limitations, the model files are hosted externally. Download them from:

1. [model-00001-of-00002.safetensors](https://huggingface.co/HibernatesAI/Hibernates-2B-R1-V1/blob/main/model-00001-of-00002.safetensors)
2. [model-00002-of-00002.safetensors](https://huggingface.co/HibernatesAI/Hibernates-2B-R1-V1/blob/main/model-00002-of-00002.safetensors)

Place these files in the root directory of the project before running.