Cluster 449609

Pre-trained transformer models and libraries for text classification tasks using RoBERTa and similar architectures. The cluster centers on fine-tuned models for sentiment analysis, emotion detection, topic classification, hate speech detection, and emoji prediction on social media text, built with PyTorch and TensorFlow. Repositories here provide both the model implementations compatible with the Hugging Face transformers ecosystem and the underlying frameworks for training and deploying these classifiers.

Infrastructure and models for text classification, semantic embeddings, and sequence labeling built on Hugging Face transformers. The cluster centers on pre-trained models optimized for inference (via safetensors and text-embeddings-inference), with particular emphasis on RoBERTa-based classifiers and smaller language model variants. Repositories range from specialized task models like citation polarity detection and medical QA reward models to general-purpose embedding services and model deployment endpoints.