Ah, I get it.
All compiling process starts with relay will call into relay.build(...), which will go through what I called the "normal build flow that starts with high-level optimizations. The process is followed up with low level optimizations mainly at TOPI level. The VTA calls vta.build() started with TOPI level optimization, thus it started with lower level optimization directly w/o the relay paths above. It is recommended we can modify the python files names, maybe. **tvm/python/tvm/relay/build_module.py** --> build_relay_modules.py **tvm/python/tvm/driver/build_module.py** --> build_topi_modules.py I now marked it as solved. Please correct me if there's anything wrong. --- [Visit Topic](https://discuss.tvm.ai/t/resolved-why-does-vta-build-has-its-own-path/6259/3) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/unsubscribe/5d7c5d71ba97275130818920b537413ae247a5488b6bb4d9051eb6192535f59a).
