[Web] Quantized model decreases in size, but takes same amount of inference time as non-quantized model

[Web] Quantized model decreases in size, but takes same amount of inference time as non-quantized model #5155