tty

分享海报

@user_a735fb41/yyjnnn

This skill should be used when the user asks about the Transformer architecture, the "Attention Is All You Need" paper (Vaswani et al., NIPS 2017), self-attention, multi-head attention, scaled dot-product attention, positional encoding, or encoder-decoder attention. Also trigger on requests to explain, summarize, compare, implement, or apply the Transformer model, or to answer questions whose answer depends on the paper architecture, equations, hyperparameters, or results.