representational capacity
Representational capacity is the ability of a model to capture and represent complex patterns in the data, dictated by factors such as model size, architecture, and depth. Higher capacity can lead to improved performance but may also risk overfitting.
- AuroRA: Breaking Low-Rank Bottleneck of LoRA with Nonlinear Mapping
- Diversity Is All You Need for Contrastive Learning: Spectral Bounds on Gradient Magnitudes
- Efficient Training of Minimal and Maximal Low-Rank Recurrent Neural Networks
- GraLoRA: Granular Low-Rank Adaptation for Parameter-Efficient Fine-Tuning
- Object-Centric Representation Learning for Enhanced 3D Semantic Scene Graph Prediction
- Pool Me Wisely: On the Effect of Pooling in Transformer-Based Models
- UniTok: a Unified Tokenizer for Visual Generation and Understanding