Public GPU and NPU uploads

Find efficient local LLM hardware per watt

Compare GPUs and NPUs by token generation per watt to find local LLM hardware that balances speed and lower power consumption.