david/aphrodite-engine: PygmalionAI's large-scale inference engine pygmalion.chat It is designed to serve as the inference endpoint for the PygmalionAI website, and to allow serving the Pygmalion models to a large number of users with blazing fast speeds (thanks to vLLM's Paged Attention).

réplica de https://github.com/PygmalionAI/aphrodite-engine

AlpinDale 9bdf8d5bfa mamba: enable continuous batching for mamba kernels (#1055)		hai 1 semana
..
causal_conv1d	9bdf8d5bfa mamba: enable continuous batching for mamba kernels (#1055)	hai 1 semana
mamba_ssm	9bdf8d5bfa mamba: enable continuous batching for mamba kernels (#1055)	hai 1 semana