# INT8 Quantization - Code generation for backends

**URL:** <https://discuss.tvm.apache.org/t/int8-quantization-code-generation-for-backends/518>\
**Category:** RFC\
**Created:** [July 26, 2018, 1:08am UTC](https://discuss.tvm.apache.org/t/int8-quantization-code-generation-for-backends/518 "2018-07-26T01:08:20Z")\
**Posts on this page:** 1\
**Showing post:** 4

<div class="post-metadata">

**Author:** ![thierry](https://discuss.tvm.apache.org/user_avatar/discuss.tvm.apache.org/thierry/32/116_2.png) [@thierry](https://discuss.tvm.apache.org/u/thierry)\
**Post date:** [July 26, 2018, 3:36am UTC](https://discuss.tvm.apache.org/t/int8-quantization-code-generation-for-backends/518/4 "2018-07-26T03:36:47Z")

</div>

Similar observation as in your other post - it might be good to think about supporting arbitrary precision operators at the graph level since there is work on supporting bit-serial convolutions on RPi, and also quantized operators in FPGAs, accelerators etc.

The challenge with arbitrary precision is data layout, and specifically how to pack data into standard 8bit/32bit words.

I’m happy to discuss some ideas, and present some concrete scenarios down the road as we continue our work on quantization with VTA.

---

_[View the full topic](https://discuss.tvm.apache.org/t/int8-quantization-code-generation-for-backends/518)._