Replace WeightOnlyInt8Linear with TorchAO int8_weight_only quantization #3328

Job	Run time
compile-gguf (macos-14)	3m 55s
gather-models-cpu	13s
gather-models-gpu	13s
runner-aoti (16-core-ubuntu)	6m 0s
runner-et (16-core-ubuntu)	10m 1s
runner-aoti (macos-14-xlarge)	3m 18s
runner-et (macos-14-xlarge)	8m 2s
test-gguf-util (macos-14)	6m 39s
test-mps (macos-m1-stable)	1m 12s
test-mps-dtype (macos-m1-stable)	1m 57s
test-tinystories-executorch (16-core-ubuntu)	17m 40s
torchchat-command-load-test (macos-14)	1m 32s
test-torchao-experimental (macos-14-xlarge)	12m 40s
test-tinystories-executorch (macos-14-xlarge)	16m 41s
test-build-runner-et-android / linux-job	13m 7s
test-cpu-compile (x86_64, stories15M)	16m 8s
test-cpu-aoti (x86_64, stories15M)	17m 55s
test-cpu-compile (aarch64, stories15M)	8m 14s
test-cpu-eval-sanity-check (x86_64, stories15M)	10m 28s
test-cpu-aoti (aarch64, stories15M)	10m 35s
test-cpu-eval-sanity-check (aarch64, stories15M)	11m 21s
test-cpu-eval-sanity-check-float16 (x86_64, stories15M)	30m 48s
test-cpu-eval-sanity-check-float32 (x86_64, stories15M)	38m 21s
test-cpu-eval-sanity-check-float16 (aarch64, stories15M)	7m 0s
test-cpu-eval-sanity-check-float32 (aarch64, stories15M)	8m 51s
test-gpu-aoti-bfloat16 (cuda, stories15M) / linux-job	14m 5s
test-gpu-aoti-float16 (cuda, stories15M) / linux-job	13m 12s
test-gpu-aoti-float32 (cuda, stories15M) / linux-job	13m 17s
test-gpu-compile (cuda, stories15M) / linux-job	23m 54s
test-gpu-eval-sanity-check (cuda, stories15M) / linux-job	12m 49s
	5h 40m 8s

Provide feedback