Hi, I tried the pyxir/examples/tvm/, which shows at the tvm doc, and I almost succeeded. But when I ran the program on the ultra96 board, I got the error.  I checked the dpu version on the host and the board, I found they were both v1.4.1.
I think my situation is really like this [issue](https://forums.xilinx.com/t5/AI-and-Vitis-AI/DNNDK-DPU-configuration-mismatch-for-kernel-xp0-on-ultra96v2/m-p/1200511), but I don't know how to fix it. I know using the vai_c_ to compile the model needs an 'arch.json', I guess lacking the file for ultra96 may cause the isssue, but I don't know how tvm uses it. By the way, I didn't use the PYNQ2.6-Z2 image for ultra96v2, which the tvm doc says, I used PYNQ2.6-Ultra96V2, I think they were different boards, and I didn't boot successfully using the PYNQ2.6-Z2 image when I try. I would appreciate it very much, if you can help me set up the flow for ultra96 v2 board. I think the doc is not very clear, and maybe it has some inconsistence with the version type among the vitis-ai. You may check [this](https://github.com/Xilinx/pyxir/issues/39) for the details, because I can't put many links and images here as a new user. --- [Visit Topic](https://discuss.tvm.apache.org/t/how-to-use-vitis-ai-integration-for-ultra96-v2-board/9883/1) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.apache.org/email/unsubscribe/e5e4de38e74624cff0c89d607f48385352aa3c49383d0c76fe5dbeb52a6b39d8).
