nthu-pllab / RISCV-DLR

This repository is a lightweight runtime for RISC-V utility.
10 stars 8 forks source link

example make error #2

Closed alexhilaev closed 3 years ago

alexhilaev commented 3 years ago

when i try to "make" an example, it gives me Traceback (most recent call last): File "build_module.py", line 5, in <module> import tvm ImportError: No module named tvm

JarvisRu commented 3 years ago

Hi @alexhilaev! Thank you for asking. As this project is based on TVM, the first thing you need to do is to make sure your env has TVM (we used v0.6) for compiling and optimizing AI models. For more details, please refer to TVM directly. And for the real showcase with this runtime, since we haven't submitted our source code of TVM that optimizes QNN for RISC-V with P extension, you can't run it directly with our sample code. It's only for demonstrating how we run and optimize for the entire flow. But if you only want to try this runtime with a simple AI model with TVM to RISC-V (without codegen with special extension - V, P extension). We may be able to generate a sample code if you need.

alexhilaev commented 3 years ago

yeah, i understand now. i`m closing it, sorry for disturbing.