TVM kernel supports dynamic shape, while rank of the shape has to be fixed. We did some experimental work before, to exhaust all the combination of shape rank + op attribute ahead of time and compile to .so. It's doable but has some restriction.
--- [Visit Topic](https://discuss.tvm.ai/t/deployment-to-pytorch-dlpack/6069/6) 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/54d79938edd1e47e7e21af4c3a7a144497e3e2f8f7cb0123067f3accae3a1c84).