mamalito345/ANKA-50M-RMW3
A 48.9M-parameter English causal language model trained from scratch on 5B tokens, featuring recurrent weight sharing, Reflect corrections, and sparse Morph routing.
正在读取项目信息…
A 48.9M-parameter English causal language model trained from scratch on 5B tokens, featuring recurrent weight sharing, Reflect corrections, and sparse Morph routing.