---

Fine-Tuning and Inference with Hugging Face Transformers

This project demonstrates how to fine-tune a pre-trained DistilBERT model on a sentiment analysis task using the IMDb dataset, and how to run inference on new data using the fine-tuned model. The project uses the Hugging Face transformers library for model fine-tuning and inference.

Project Structure

├── results/                  # Directory for saving finetuned model
├── finetune.py               # Main script for fine-tuning the model
├── inference.py              # running inference with the fine-tuned model
├── infer_base.py             # running inference with the base model
├── .gitignore                
└── README.md                 

Prerequisites

• Python 3.6+
• transformers library
• datasets library
• PyTorch
• CUDA-enabled GPU (optional but recommended for faster training)

You can install the dependencies using the following command:

pip install torch transformers datasets

Steps to Run the Project

1. Fine-Tune the Model

Use the finetune.py script to fine-tune the DistilBERT model on the IMDb dataset. The script fine-tunes the model and saves it under the ./results directory.

python finetune.py

2. Save the Fine-Tuned Model

The fine-tuned model is saved automatically at the end of training in the results/ directory.

3. Run Inference

Once the model is fine-tuned, use the inference.py script to run inference on new input texts. The script loads the fine-tuned model and tokenizer and prints the predicted sentiment (positive/negative) for each input.

python inference.py

Example Output

Input: This movie was fantastic! I loved every minute of it.
Predicted Sentiment: Positive

Input: I didn't like this movie. It was really boring and poorly made.
Predicted Sentiment: Negative

Project Files

1. finetune.py

• Description: This script fine-tunes a pre-trained DistilBERT model on the IMDb dataset. It tokenizes the dataset, configures the model and trainer, and saves the fine-tuned model at the end of training.

2. inference.py

• Description: This script loads the fine-tuned model and runs inference on sample text inputs.

3. infer_base.py

• Description: This script loads the base DistilBERT model and runs inference on sample text inputs.

License

This project is licensed under the MIT License. See the LICENSE file for details.

---