Alex Eftimiades

I tried make clean and make all and got the same error. Then I tried export CLUSTER_NAME=k8s and got this error: Failed to parse input: unexpected end of JSON input...

I deleted all buckets, but terraform had a lock on it. I tried make clean, and terraform destroy, then deleting the the tfstate file and again deleted s3 buckets. This...

Oh my understanding was xgboost has a separate split for whether the data is null (e.g. is of a separate "null type", and if not, proceed with dealing with it...

Thanks for the link and I really appreciate you looking into the feature request! The gist of the the modification would be associate null values with both branches but weight...

[It looks like](https://stackoverflow.com/questions/68327863/importing-jax-fails-on-mac-with-m1-chip) for python3.9.6, the latest jax/jaxlib support Apple's new M1 chips. Let me know if switching to that version solves the problem

When I try ``` from jax import export class MyClass(eqx.Module): @eqx.filter_jit def fn(self, ...): .... model = MyClass(...) exported = export.export(model.fn._cached) exported(*args, **kwargs).serialize() ``` I get this error. ``` Non-hashable...

Just following up with the minimal example to replicate the problem. ``` class Test(eqx.Module): @eqx.filter_jit def fn(self, data): return data obj = Test() fn = obj.fn x = jnp.array(3.0) fn(x)...