Fix issue with device check failure

[jonb377]

Fixes / Features

• If device type check fails both GKE and accelerator type and cluster device type, the workload creation will succeed anyway. Such workloads should not be created, since there can be incompatibilities between the workload and cluster configs.

Testing / Documentation

Attempted to create a workload with invalid device type for the cluster:

(xpk) cloudtop [~/xpk] % python xpk.py workload create --device-type h100-80gb-8 --project XXXXXX --zone XXXXXX --cluster XXXXXX --docker-image XXXXXX \
                --command "echo hello" \
                --num-nodes 2 --workload jonbolin-test-$RANDOM
[XPK] Starting xpk
...
[XPK] Gke Accelerator Type Check: jonbolin-test-998 is requesting nvidia-h100-80gb but cluster only contains dict_keys(['h100-mega-80gb-8']).
[XPK] Device Type Check: jonbolin-test-aot-998 is requesting h100-80gb-8 but cluster only contains dict_keys(['h100-mega-80gb-8']).
[XPK] Both Device Type and GKE Accelerator Type checks failed. XPK will not create the workload jonbolin-test-aot-998.
[XPK] XPK failed, error code 1

• [ y ] Tests pass
• [ y ] Appropriate changes to documentation are included in the PR

[jonb377]

@Obliviour It seems the tests are failing, but it seems unrelated. Should I retry or can we override?