Scvi model kills kernel when .train() is called on Macbook Pro with Apple M1 Pro Chip

Help scvi-tools
1

After installing scvi-tools on Macbook Pro with Apple M1 Pro Chip, .train() automatically kills the kernel in jupyter notebook. I have followed the standard procedure on even the smallest data set and received this error

Intel MKL FATAL ERROR: This system does not meet the minimum requirements for use of the Intel(R) Math Kernel Library.
The processor must support the Intel(R) Supplemental Streaming SIMD Extensions 3 (Intel(R) SSSE3) instructions.
The processor must support the Intel(R) Streaming SIMD Extensions 4.2 (Intel(R) SSE4.2) instructions.
The processor must support the Intel(R) Advanced Vector Extensions (Intel(R) AVX) instructions.

Is there something I am doing wrong?

Please help!

2

Same thing happened when I installed with my existing conda setup. Truthfully don’t remember if that installation was native or via Rosetta, but I deleted it and installed the arm64 mambforge build from here as recommended in the scvi-tools installation page. No issues now.

3

This! scvi-tools won’t work on apple silicon without a native python version

4

Thanks for your reply! I did realize this and used the recommended build. Works good now.