Rebase community TensorFlow. (#1213)

* Rebased community TensorFlow.

* Fix rebase error of tf_community.

* Fix BUILD and docker env.

* Fix build error.

* Disable remote cache for testing.

* Update CUDA version of CI.

* Migrate T4 to A10 for CI.

* Clean CI.

* Refine code.

* Update document.

* Update paper authors.

* Fix rebase errors.

* Add missed tf250 requirement file.

* Update torch-pre version.

* Update torch-pre version.

* Update torch-pre dependency version.

* Update torch-pre-cpu version.

* Fix dependency error of tensorflow-blade.

* Fixdeps for AArch64.

* Update bazel deps of tensorflow_blade.

* Fix transform-dialect errors.

* Update tf250 CI file.

* Temporarily disabled some UTs.

* Temporarily disable remote cache for testing.

* Enable CI remote cache and add python version for tf250_cpu.

* Update scaled_dot_product_efficient_attention prototype.

* Change pdl-type to transform-type.

* Disable the buggy AArch64 schedules, and add DCE, CSE, LICM Transform-Dialect ops.

* Update README.

* Update README.

* Update README.

* Update tensorflow blade workspace.

Author: JamesTheZ

.bazelversion

@@ -1 +1 @@
-5.1.1
+6.1.0

.github/workflows/d2_trt_bench.yml

@@ -15,7 +15,7 @@ jobs:
     with:
       name: torch-trt-and-d2-benchmark
       base_image: nvcr.io/nvidia/pytorch:22.01-py3
-      device: gpu-t4
+      device: gpu-a10
       dockerfile: docker/cronjobs/Dockerfile.torch.ngc
       extra_run_args: --gpus all
       exec_command: bash -c "cd pytorch_blade && bash ./benchmark/run_benchmark.sh"