Question) Is there a way to control the weight precision in CONV tutorial?

[a2jinhee]

Hello, thank you for sharing this great work! I’m new to Timeloop and currently working through the tutorial exercises. I’m wondering if there’s a way to adjust the weight precision in the problem space. For example, in exercise 06-mapper-convlayer-eyeriss, instead of loading FP32 weights into the SRAM, could we modify it to load weights in INT8 format and then calculate the power consumption for that scenario? If this is possible, could you give me some pointers on what to change?