Revert "Fix --bf16 option support for Neuron after PR #22300" (#22451)

This reverts commit fd81746dbec5f17c8285a0fdc72ca4b4c025cc33.

src/transformers/trainer.py

@@ -588,12 +588,7 @@ class Trainer:
 
         if args.fp16 or args.bf16:
             if args.half_precision_backend == "auto":
-                if is_torch_neuroncore_available():
-                    if args.fp16:
-                        raise ValueError("Tried to use `fp16` but this option is not yet supported on Neuron.")
-                    else:
-                        args.half_precision_backend = "cpu_amp"
-                elif args.device == torch.device("cpu"):
+                if args.device == torch.device("cpu"):
                     if args.fp16:
                         raise ValueError("Tried to use `fp16` but it is not supported on cpu")
                     elif _is_native_cpu_amp_available: