20) Mixture of Experts Balancing Techniques Auxiliary Loss Load Balancing Capacity Factor5просмотров2 месяца назад
15) All about Sinusoidal Positional Encodings What’s with the weird sin-cos formula1просмотр2 месяца назад
14) Integer and Binary Positional Encodings Journey towards Rotary Positional Encodings (RoPE)5просмотров2 месяца назад
12) Multi-Head Latent Attention From Scratch One of the major DeepSeek innovation8просмотров2 месяца назад
11) Understand Grouped Query Attention (GQA) The final frontier before latent attention8просмотров2 месяца назад
10) Multi-Query Attention Explained Dealing with KV Cache Memory Issues Part 19просмотров2 месяца назад