Error: "Feature "bf16" requires .target sm_80 or higher

[7wik3vedi16]

Error: "Feature "bf16" requires .target sm_80 or higher

Is there any other way other than upgrading my GPU ?

[conglu1997]

Hey! This isn't necessary for performance, you could convert these to full precision floats! :)

[xprabhudayal]

@7wik3vedi16 you can try changing the dtype = float16 from the experiment.py, but that would be really slower as compared with bfloat16 ive tried it in google colab and same error happened, maybe you can try with a100/h100 gpu