Tagged: few-shot-learning

9 articles

Latest

GPT-Neo needs just 3 examples to match GPT-3's few-shot trick

SetFit hits 98% of T-Few's accuracy with 30x fewer parameters and no prompts

SetFit Cut Witty Works' Training Data Needs to 20 Sentences Per Word

Intel and Hugging Face's SetFitABSA beats Llama2 in few-shot sentiment analysis

Reptile: OpenAI's meta-learning algo skips the heavy math

Reptile: Meta-Learning's Shortcut Drops Second-Order Math, Still Works

DeepMind's New Model Learns Spatial Concepts from Just 5 Examples

GPT-3: 175B params, zero fine-tuning, and humans can't tell

SetFit Inference Hits 7.8x Speedup on Intel Xeon With One Quantization Step