Less Trainable Parameters after quantization

Hello, @dfrank.

Looking for an answer to the same question. Along with what you have mentioned above, the no. of total parameters reduces as well post-quantization. Please let me know if you figured out why.

1 Like