Post

Unfortunately, being GPU-poor means that training models even as big as LFM2.5-2.6B takes ages. 🤞

0 reposts0 likes

No replies yet.