Announcing end of support for the Torch/XLA inference path on Trn2 and Trn3 starting in Neuron 2.34

As part of the transition to native PyTorch support on Trainium, native PyTorch support through TorchNeuron is now available starting with Neuron Release 2.33. With native PyTorch support in place, the Torch/XLA inference path is no longer required on Trn2 and Trn3.

Starting with the next full Neuron release, Neuron will end support for the Torch/XLA inference path on Trn2 and Trn3 instances. The native PyTorch path remains the default and recommended path.

On Inf2 and Trn1 instances, Torch/XLA inference support remains in maintenance mode, as described in Torch/XLA support for Inference on Inf2 and Trn1 instances is now in maintenance mode starting with Neuron 2.31.