This project is all about showing how one can fine-tune models in a simple, straightforward way. Lightweight fine-tuning is one of the most important techniques for adapting foundation models, because it allows you to modify foundation models for your needs without needing substantial computational resources.
In this project, we will bring together all of the essential components of a PyTorch + Hugging Face training and inference process. Specifically, we will:
- Load a pre-trained model and evaluate its performance
- Perform parameter-efficient fine tuning using the pre-trained model
- Perform inference using the fine-tuned model and compare its performance to the original model
The dataset consists of two columns:
- tweet: where each line includes the text of a tweet that included emoji (but the emoji has been removed).
- emoji: where each line includes the name of the emoji for the corresponding text in tweet.
- Is Fine-Tuning Necessary?
- Predictive Modeling: We used transformers to train our dataset and make it predict tweets or text by giving certain Emojis suitable for that given text.
- Analysis On PEFT
For a detailed look into the analysis, check out this Jupyter Notebook