MHA2MLA
Collection
The MHA2MLA model published in the paper "Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-Based LLMs" • 22 items • Updated • 2
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.