Stars
Efficient Python library for Extended LSTM with exponential gating, memory mixing, and matrix memory for superior sequence modeling.
RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RN…
wengong-jin / chemprop
Forked from chemprop/chempropChemical Property Prediction with Graph Convolutional Networks