[Distributed] create model on meta device #3025

Job	Run time
compile-gguf (macos-14)	4m 25s
gather-models-cpu	12s
gather-models-gpu	14s
runner-aoti (16-core-ubuntu)	4m 40s
runner-et (16-core-ubuntu)	6m 26s
runner-aoti (macos-14-xlarge)	3m 8s
runner-et (macos-14-xlarge)	5m 3s
test-gguf-util (macos-14)	4m 32s
test-mps (macos-m1-stable)	1m 55s
test-mps-dtype (macos-m1-stable)	2m 30s
test-tinystories-executorch (16-core-ubuntu)	19m 41s
test-torchao-experimental (macos-14-xlarge)	5m 36s
test-tinystories-executorch (macos-14-xlarge)	17m 25s
torchchat-command-load-test (macos-14)	1m 44s
test-build-runner-et-android / linux-job	12m 16s
test-cpu-compile (x86_64, stories15M)	16m 32s
test-cpu-eval-sanity-check (x86_64, stories15M)	9m 13s
test-cpu-compile (aarch64, stories15M)	8m 47s
test-cpu-eval-sanity-check-float16 (x86_64, stories15M)	29m 8s
test-cpu-eval-sanity-check (aarch64, stories15M)	8m 26s
test-cpu-aoti (x86_64, stories15M)	17m 53s
test-cpu-eval-sanity-check-float16 (aarch64, stories15M)	6m 43s
test-cpu-aoti (aarch64, stories15M)	9m 51s
test-cpu-eval-sanity-check-float32 (x86_64, stories15M)	36m 18s
test-cpu-eval-sanity-check-float32 (aarch64, stories15M)	8m 12s
test-gpu-aoti-bfloat16 (cuda, stories15M) / linux-job	13m 19s
test-gpu-aoti-float32 (cuda, stories15M) / linux-job	12m 15s
test-gpu-aoti-float16 (cuda, stories15M) / linux-job	12m 13s
test-gpu-compile (cuda, stories15M) / linux-job	23m 32s
test-gpu-eval-sanity-check (cuda, stories15M) / linux-job	11m 24s
	5h 13m 33s

Provide feedback