BlinkDL/RWKV-LM
RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RNN and transformer - great performance, linear time, constant space (no kv-cache), fast training, infinite ctx_len, and free sentence embedding.
PythonApache-2.0
Watchers
- altaic
- bmorphism@plurigrid
- bradjonescaNew York
- brataoEscavador
- celsopitta
- desktable
- dsoul
- edwardxieChina
- flyinweb
- hahattMicrosoft
- haroldyongChina, Hangzhou
- Jack47Sensetime
- jhcloos
- JohnnyOpcodeToronto, Ontario, Canada
- kahkeng
- kreelyAdaptable Solutions
- kziemskiQCOMPUTE
- martyn255LABS
- mbofb
- michalwolsNew York
- mklySan Francisco, CA
- photosbanIntel Corporation
- python33
- QubitiumModelCloud.ai
- scorpioliu
- seshakiran
- sterling312
- tiendung
- trappedinspacetimeFor Personal Use
- trycatcher
- vizantoThe Earth
- WilliamMahler
- wookayinUniversity of Michigan
- wqlxx
- zebrajackpungke
- zhangyijunShanghai