ai-by-hand-deepseek-solution
Public? AI-by-hand: Multi-head Latent Attention, RoPE, and MoE in Deepseek.
Creat:2025-02-01T08:33:49
Update:2025-02-17T09:09:15
2
Stars
0
Stars Increase
? AI-by-hand: Multi-head Latent Attention, RoPE, and MoE in Deepseek.