Pinned Loading

1. Contrastive-Learning-Based-Text-Classification Contrastive-Learning-Based-Text-Classification Public

   A Supercharged two-phase transformers based text classification training pipeline with: contrastive pre-training + asymmetric loss. Perfect for imbalanced datasets! 🚀

   Jupyter Notebook

2. Deep-Learning-Image-Captioning Deep-Learning-Image-Captioning Public

   An end-to-end deep learning solution for automatic image captioning using convolutional neural networks (CNN) as encoders and recurrent neural networks (RNN) as decoders in PyTorch.

   Jupyter Notebook

3. HelpSteer-ComplexityPrediction HelpSteer-ComplexityPrediction Public

   Predict the complexity of AI-generated responses using a fine-tuned transformer model trained on NVIDIA's HelpSteer dataset using an asymmetric quantile loss function for imbalanced data.

   Jupyter Notebook