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). @ sampling-kernels

AlpinDale cac8163f77 fix: request abort crashing pipeline parallel		5 月之前
..
__init__.py	9d81716bfd [v0.5.3] Release Candidate (#388)	8 月之前
interfaces.py	ae04f57ec1 feat: Pipeline Parallel support (#581)	5 月之前
multi_step.py	ae04f57ec1 feat: Pipeline Parallel support (#581)	5 月之前
single_step.py	cac8163f77 fix: request abort crashing pipeline parallel	5 月之前
stop_checker.py	6785d78d82 fix: do not expose EOS token in the API	5 月之前
util.py	90ceab32ff refactor: consolidate prompt args to LLM engines	5 月之前