search

thunlp / InfLLM

The code of our paper "InfLLM: Unveiling the Intrinsic Capacity of LLMs for Understanding Extremely Long Sequences with Training-Free Memory"
MIT License
269 stars 21 forks source link

Optim memory usage, fastchat integration and multiprocessing benchmark #13

Closed guyan364 closed 5 months ago

guyan364 commented 5 months ago
  1. 优化显存使用
  2. 整合 FastChat chat cli

  3. 多进程 benchmark

  4. Optimize memory usage
  5. Integrate FastChat chat CLI
  6. Multi-process benchmarking

close #5 #8 #11 #12