Update PyTorch pin and enable MPS qops #1410

Job	Run time
compile-gguf (macos-14)	6m 6s
gather-models-cpu	12s
gather-models-gpu	13s
runner-aoti (16-core-ubuntu)	7m 46s
runner-aoti (macos-14-xlarge)	4m 45s
runner-et (16-core-ubuntu)	4m 43s
test-gguf-util (macos-14)	10m 38s
runner-et (macos-14-xlarge)	4m 41s
test-tinystories-executorch (16-core-ubuntu)	17m 43s
torchchat-command-load-test (macos-14)	2m 34s
test-tinystories-executorch (macos-14-xlarge)	16m 58s
test-mps / macos-job	1m 13s
test-mps-dtype / macos-job	1m 12s
test-build-runner-et-android / linux-job	15m 27s
test-cpu-eval-sanity-check (x86_64, stories15M)	9m 54s
test-cpu-aoti (x86_64, stories15M)	21m 41s
test-cpu-eval-sanity-check-float16 (x86_64, stories15M)	19m 55s
test-cpu-eval-sanity-check (aarch64, stories15M)	28m 59s
test-cpu-compile (x86_64, stories15M)	12m 2s
test-cpu-aoti (aarch64, stories15M)	9m 50s
test-cpu-eval-sanity-check-float16 (aarch64, stories15M)	5m 49s
test-cpu-compile (aarch64, stories15M)	7m 56s
test-cpu-eval-sanity-check-float32 (x86_64, stories15M)	26m 41s
test-cpu-eval-sanity-check-float32 (aarch64, stories15M)	27m 43s
test-gpu-aoti-bfloat16 (cuda, stories15M) / linux-job	6m 49s
test-gpu-aoti-float16 (cuda, stories15M) / linux-job	6m 45s
test-gpu-aoti-float32 (cuda, stories15M) / linux-job	6m 47s
test-gpu-compile (cuda, stories15M) / linux-job	6m 44s
test-gpu-eval-sanity-check (cuda, stories15M) / linux-job	8m 29s
	5h 0m 15s

Provide feedback