[static runtime] add more _out variants (#48260)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48260
supporting a couple more operators
Test Plan:
use Ansha's test framework for e2e test
```
numactl -m 0 -C 3 ./buck-out/opt/gen/caffe2/caffe2/fb/predictor/ptvsc2_predictor_bench --pred_net=/home/bwasti/adindexer/precomputation_merge_net.pb --c2_inputs=/home/bwasti/adindexer/c2_inputs_precomputation_bs1.pb --c2_weights=/home/bwasti/adindexer/c2_weights_precomputation.pb --scripted_model=/home/bwasti/adindexer/traced_precomputation_partial_dper_fixes.pt --pt_inputs=/home/bwasti/adindexer/container_precomputation_bs1.pt --iters=30000 --warmup_iters=10000 --num_threads=1 --pt_enable_static_runtime=true --pt_cleanup_activations=true --pt_enable_out_variant=true --eps 1e-2
```
Reviewed By: hlu1
Differential Revision: D24767322
fbshipit-source-id: dce7f9bc0427632129f263bad509f0f00a21ccf3