BlinkDL/RWKV-LM
RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.
PythonApache-2.0
Stargazers
- akzaidi@Microsoft
- back2yes
- BentengMa
- BlinkDLhttp://withablink.com
- brataoEscavador
- cfoster0
- CrustaceanJ
- Cuda-ChenSeeking for opportunities
- davda54LTG, UiO
- davidnvqJapan
- dnbaker@langmead-lab
- edwin-19Malaysia
- efiLeipzig University
- flowpoint
- fly51flyPRIS
- hypnopump
- igorbrigadirInsight Centre for Data Analytics
- jon-chunKenyon College
- JunnYu@ECUST
- leeyangUniversity of Michigan
- lgstd
- likaihereBejing China
- lucidrainsSan Francisco
- MicPieOpenBioML.org
- napoler
- neverix
- polyrandBarcelona
- sailfish009freelancer
- sdtblck
- SSshuishui
- StellaAthenaBooz Allen Hamilton, EleutherAI
- theblackcat102iKala
- TheodoreGalanosAustrian Institute of Technology
- vicgalleKomorebi AI & ICMAT-CSIC
- wx-bRIOS
- zirui-yuan