@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 17, 09:15 UTC

main

Merge pre_step and step

Woosuk Kwon committed over 3 years ago

Reduce the number of states in scheduler

Woosuk Kwon committed over 3 years ago

Refactor and annotate types for attention

Woosuk Kwon committed over 3 years ago

Remove xformers

Woosuk Kwon committed over 3 years ago

[WIP] Add server script

Woosuk Kwon committed over 3 years ago

Fix attention

Woosuk Kwon committed over 3 years ago

Fix attention

Woosuk Kwon committed over 3 years ago

Fix attention

Woosuk Kwon committed over 3 years ago

Set default dtype to half

Woosuk Kwon committed over 3 years ago

Fix sampler

Woosuk Kwon committed over 3 years ago

Minor

Woosuk Kwon committed over 3 years ago

Consider pempty tensor

Woosuk Kwon committed over 3 years ago

Fix typo

Woosuk Kwon committed over 3 years ago

Add controller

Woosuk Kwon committed over 3 years ago

Add unoptimized OPT Attention

Woosuk Kwon committed over 3 years ago

Add greedy sampler

Woosuk Kwon committed over 3 years ago

Add seq_ids to input metadata

Woosuk Kwon committed over 3 years ago

Add get_block_table

Woosuk Kwon committed over 3 years ago

decoding.py -> sampling_params.py

Woosuk Kwon committed over 3 years ago

← Newer Page 1117 of 1120 Older →