user@devops:~$ cat README.md
Fine-tuning Trainer
# Description
Transformer model fine-tuning project using Hugging Face Trainer API. Implements BERT, RoBERTa and DistilBERT fine-tuning for text classification tasks with custom datasets (CSV/JSON). Covers the full pipeline: dataset loading and preprocessing, tokenization with label mapping, TrainingArguments configuration (learning rate, batch size, epochs, logging, evaluation), Trainer training, evaluation with custom metrics (accuracy, F1, precision, recall), fine-tuned model saving and upload to Hugging Face Hub. Includes performance comparison between base and fine-tuned models.
# Key features
$ Fine-tuning BERT, RoBERTa and DistilBERT for classification
$ Complete Trainer API: TrainingArguments, metrics, callbacks
$ Custom dataset loading from CSV/JSON
$ Tokenization with label mapping and truncation/padding
$ Evaluation with accuracy, F1, precision and recall
$ Fine-tuned model upload to Hugging Face Hub
# Gallery
# Technologies used
