What do you think? It can't compete with NumPy in performance.
Nouse doesn't compete with NumPy — they're in completely
different categories. NumPy is numerical array computation.
Nouse is a knowledge graph that injects structured context
into LLM prompts.
The relevant comparison is: does a small model with Nouse
answer domain questions better than a large model without it?
Benchmark says yes. 96% vs 47%.
Yes, it was javascript-based package. NumPy is highly optimized by Native. It can't be competed. But starlight-numera is still useful for advanced tasks like computing and data science - related tasks on web and scripting.