Commits · b7955ef17b8d899327b25564f20665ec3ffa71cb · norm / vllm

19 May, 2023 2 commits
- Fix timeout error in the FastAPI frontend (#34) · b7955ef1
  Zhuohan Li authored May 19, 2023
  
  b7955ef1
- Use runtime profiling to replace manual memory analyzers (#81) · f756799b
  Zhuohan Li authored May 19, 2023
  
  f756799b
11 May, 2023 2 commits
- Enhance SamplingParams (#96) · 42f1042e
  Woosuk Kwon authored May 11, 2023
  
  42f1042e
- Implement presence and frequency penalties (#95) · 55f8b0a5
  Woosuk Kwon authored May 10, 2023
  
  55f8b0a5
09 May, 2023 2 commits
- Use slow tokenizer for LLaMA (#84) · 85eb6318
  Woosuk Kwon authored May 09, 2023
  
  85eb6318
- Refactor system architecture (#82) · 7c041ab5
  Woosuk Kwon authored May 09, 2023
  
  7c041ab5
03 May, 2023 1 commit
- New weight loader without np copy (#52) · 27f1410d
  Zhuohan Li authored May 03, 2023
  
  27f1410d
30 Apr, 2023 1 commit
- Add an option to launch cacheflow without ray (#51) · 4858f3bb
  Zhuohan Li authored Apr 30, 2023
  
  4858f3bb
09 Apr, 2023 1 commit
- Add an option to use dummy model weights (#33) · ee88a7e5
  Woosuk Kwon authored Apr 08, 2023
  
  ee88a7e5
06 Apr, 2023 1 commit
- Fix potential bugs in FastAPI frontend and add comments (#28) · a490aafa
  Zhuohan Li authored Apr 06, 2023
  
  a490aafa
30 Mar, 2023 1 commit
- Implement preemption via recomputation & Refactor scheduling logic (#12) · 7a7929ab
  Woosuk Kwon authored Mar 30, 2023
  
  7a7929ab
29 Mar, 2023 1 commit
- FastAPI-based working frontend (#10) · 721fa3df
  Zhuohan Li authored Mar 29, 2023
  
  721fa3df