---
library_name: transformers
license: apache-2.0
datasets:
- ThinkAgents/Function-Calling-with-Chain-of-Thoughts
base_model:
- meta-llama/Llama-3.2-1B-Instruct
---

# Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->



## Model Details

### Model Description

`ThinkAgent-1B` is a fine-tuned version of `meta-llama/Llama-3.2-1B-Instruct`, specifically optimized for function calling tasks. The model is trained to improve Abstract Syntax Tree (AST) accuracy, ensuring more precise function call generations.

This model was developed as a graduation project by:
- **Ayman Tarig** ([GitHub](https://github.com/0xayman))
- **Tasneem Midhat** ([GitHub](https://github.com/tasneemmidhat10))

from the **University of Khartoum, Electrical and Electronics Engineering Department**.

For future work, further alignment of the model could be explored to enhance its function-calling capabilities.

## Model Details

- **Developed by:** Ayman Tarig, Tasneem Midhat
- **Model type:** Fine-tuned language model for function calling
- **Language(s) (NLP):** English
- **License:** Apache-2.0
- **Finetuned from model:** meta-llama/Llama-3.2-1B-Instruct

## Uses

### Direct Use

The model can be used as-is for function-calling tasks, providing structured outputs for function calls based on natural language prompts.

### Downstream Use

The model can be further fine-tuned for more specific function-calling applications, such as integrating into APIs or smart assistants.

### Out-of-Scope Use

The model is not intended for general-purpose conversation, creative writing, or any tasks beyond function calling and structured response generation.

## Bias, Risks, and Limitations

- The model is fine-tuned for function calling and may not generalize well outside this domain.
- The model’s outputs depend on the provided function definitions; incorrect function descriptions may lead to incorrect outputs.
- Potential biases inherited from the base model and dataset remain.

### Recommendations

Users should validate the function call outputs before execution to avoid errors or unintended consequences.


## Training Details

### Training Data

The model was fine-tuned using the **ThinkAgents/Function-Calling-with-Chain-of-Thoughts** dataset. The dataset includes reasoning chains distilled from `deepseek-ai/DeepSeek-R1-Distill-Llama-8B` to enhance function-calling accuracy.

## Model Card Authors

- Ayman Tarig ([GitHub](https://github.com/0xayman))
- Tasneem Midhat ([GitHub](https://github.com/tasneemmidhat10))