| # CEC Assistant - Llama 3.1 Fine-Tuned Model |
|
|
| ## Overview |
|
|
| CEC Assistant is a domain-specific language model fine-tuned on College of Engineering Chengannur (CEC) related information. The model is designed to answer questions about academics, departments, admissions, placements, campus facilities, student activities, and other college-related topics. |
|
|
| ## Model Details |
|
|
| - Base Model: Meta Llama 3.1 |
| - Task: Causal Language Modeling (CLM) |
| - Framework: Hugging Face Transformers |
| - Fine-Tuning Approach: Supervised Fine-Tuning |
| - Language: English |
|
|
| ## Training Process |
|
|
| The model was fine-tuned using custom college-specific data and trained using the Hugging Face Transformers ecosystem. |
|
|
| Training workflow included: |
|
|
| - Data collection and preprocessing |
| - Text cleaning and tokenization |
| - Dataset splitting for training and evaluation |
| - Causal Language Model fine-tuning |
| - Model checkpointing and validation |
| - Hugging Face model publishing |
|
|
| ## Technologies Used |
|
|
| - Transformers |
| - Datasets |
| - Trainer API |
| - AutoModelForCausalLM |
| - Llama Tokenizer |
| - PyTorch |
|
|
| ## Capabilities |
|
|
| The model can assist with: |
|
|
| - College information queries |
| - Academic program details |
| - Department-related information |
| - Admission-related questions |
| - Campus facilities and activities |
| - General institutional knowledge |
|
|
| ## Limitations |
|
|
| - Responses may not always reflect the latest institutional updates. |
| - The model may generate inaccurate information in some cases. |
| - Official college sources should be consulted for critical decisions. |
|
|
| ## License |
|
|
| This model inherits the licensing requirements of the underlying Llama 3.1 model. Users must comply with Meta's Llama license terms. |
|
|
| ## Disclaimer |
|
|
| This is an independent educational and research project and is not an official information source of College of Engineering Chengannur. |
