@kevindewalt.eth
13 ways to optimize LLM apps:
1 Right model for the task
2 Right model size/version
3 Quantization
4 Fine tuning
5 Improve prompts
6 Provide examples
7 Add context (RAG)
8 Provide structured data access
9 Integrate multiple information sources
10 Agents
11 Tools
12 Out-of-context variables
13 Orchestrate many agents
0 reply
0 recast
0 reaction