# Can anyone please explain TFLite quantization part found in Netron neural network viewer?

I was checking tflite file in Netron. There I found the quantization formula in Netron as below:

quantization: 0.007709330413490534 * (q + 3)


I know the quantization formula is: quantized_value = (r/S + Z) Where, r = input, S = scale and Z = zero-point

I have used quantization-aware training before converting the model to the TFLite model. I have uploaded the TFlite Model file to Netron App website where I have found quantization as below image. Can anyone explain What do the q and 3 mean in Netron quantization?

• Hello. Can you 1. put your specific question in the title, and 2. provide the link to the source where you took that screenshot from?
– nbro
Jan 4 at 9:37
• I have added the file and edited the title.
– Amit
Jan 4 at 13:23
• Thanks. You may also want to format the equations with latex/mathjax to improve the readability.
– nbro
Jan 4 at 13:44