موارد ضروری که برای فراگیری مدلهای LLM مورد نیاز هست:
- token embeddings
- positional embeddings
- self-attention
- transformers
- intuitive understanding of Q, K, V
- causal and multi-head attention
- temperature, top-k, top-p
- classification & instruction fine-tuning
- rotary positional encoding (RoPE)
- KV Cache
- infini-attention (long context windows)
- mixture of experts (MoE)
- grouped query attention
- basic understanding of llama-2 architecture and techniques which is actually a recap of all the previous subjects
#علوم_پزشکی #مقاله #ایده_جذاب #منابع #کلاس_آموزشی
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
🔸 @AI_Person
- token embeddings
- positional embeddings
- self-attention
- transformers
- intuitive understanding of Q, K, V
- causal and multi-head attention
- temperature, top-k, top-p
- classification & instruction fine-tuning
- rotary positional encoding (RoPE)
- KV Cache
- infini-attention (long context windows)
- mixture of experts (MoE)
- grouped query attention
- basic understanding of llama-2 architecture and techniques which is actually a recap of all the previous subjects
#علوم_پزشکی #مقاله #ایده_جذاب #منابع #کلاس_آموزشی
🔸 مطالب بیشتر 👇👇
✅ @AI_DeepMind
🔸 @AI_Person