blinkdl/rwkv-lm
steadyRWKV (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.
Python
View on GitHub
Stars
14,738
Forks
1,022
Open issues
128
24h
0
0.0%
7d
+17
+0.1%
Refresh
2h
Star history (7 days)
Last checked
14m ago
Last pushed
6d ago
Next check
just now