Use convert_to_tensor inside __init__ to avoid eager-only numpy() under XLA

.bazelrc

@@ -302,9 +302,11 @@ common:cuda --@local_config_cuda//:enable_cuda
 common:cuda --config=cuda_version
 # This flag is needed to include CUDA libraries.
 common:cuda --@local_config_cuda//cuda:include_cuda_libs=true
+common:cuda --@cuda_driver//:include_cuda_umd_libs=true
 
 # This configuration is used for building the wheels.
 common:cuda_wheel --@local_config_cuda//cuda:include_cuda_libs=false
+common:cuda_wheel --@cuda_driver//:include_cuda_umd_libs=false
 
 # CUDA: This config refers to building CUDA op kernels with clang.
 common:cuda_clang --config=cuda
@@ -596,7 +598,6 @@ common:use_tar_archive_files --repo_env=USE_LLVM_TAR_ARCHIVE_FILES=1
 common:use_tar_archive_files --repo_env=USE_MIRRORED_TAR_ARCHIVE_FILES=1
 
 # Make Bazel not try to probe the host system for a C++ toolchain.
-common:rbe_base --config=use_tar_archive_files
 common:rbe_base --config=resultstore
 common:rbe_base --repo_env=BAZEL_DO_NOT_DETECT_CPP_TOOLCHAIN=1
 common:rbe_base --define=EXECUTOR=remote
@@ -639,8 +640,8 @@ common:rbe_linux_cpu --remote_instance_name=projects/tensorflow-testing/instance
 # Download CUDA/CUDNN redistributions to preserve the repositories cache between
 # CPU and GPU builds.
 # TODO(ybaturina): Uncomment when RBE is ready to support this.
-commonld:rbe_linux_cpu --repo_env USE_CUDA_REDISTRIBUTIONS=1
-commonld:rbe_linux_cpu --config=cuda_version
+common:rbe_linux_cpu --repo_env USE_CUDA_REDISTRIBUTIONS=1
+common:rbe_linux_cpu --config=cuda_version
 
 # Deprecated RBE config with non-hermetic toolchains.
 common:rbe_linux_cpu_clang_local --config=rbe_linux_cpu
@@ -666,9 +667,6 @@ common:rbe_linux_cuda --config=cuda_clang_official
 common:rbe_linux_cuda --config=rbe_linux_cpu
 # For Remote build execution -- GPU configuration
 common:rbe_linux_cuda --repo_env=REMOTE_GPU_TESTING=1
-# Enable forward compatibility for CUDA builds because RBE docker image doesn't
-# have latest CUDA drivers installed.
-common:rbe_linux_cuda --@cuda_driver//:enable_forward_compatibility=true
 
 common:rbe_linux_cuda_nvcc --config=rbe_linux_cuda
 common:rbe_linux_cuda_nvcc --config=cuda_nvcc
@@ -861,7 +859,7 @@ test:linux_cpu_wheel_test --@local_xla//third_party/py:wheel_dependency=true --c
 test:linux_cuda_wheel_test_filters --test_tag_filters=gpu,requires-gpu,-no_gpu,-no_oss,-tf_tosa,-oss_excluded,-oss_serial,-benchmark-test,-no_cuda11,-no_oss_py38,-no_oss_py39,-no_oss_py310,-no_oss_py313
 test:linux_cuda_wheel_test_filters --build_tag_filters=gpu,requires-gpu,-no_gpu,-no_oss,-tf_tosa,-oss_excluded,-oss_serial,-benchmark-test,-no_cuda11,-no_oss_py38,-no_oss_py39,-no_oss_py310,-no_oss_py313
 test:linux_cuda_wheel_test_filters --test_lang_filters=py --test_size_filters=small,medium
-test:linux_cuda_wheel_test --@local_xla//third_party/py:wheel_dependency=true --config=linux_cuda_wheel_test_filters -- //tensorflow/... //tensorflow/tools/pip_package:prebuilt_wheel_import_api_packages_test_gpu -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
+test:linux_cuda_wheel_test --repo_env=HERMETIC_CUDA_UMD_VERSION=12.8.1 --@local_xla//third_party/py:wheel_dependency=true --config=linux_cuda_wheel_test_filters -- //tensorflow/... //tensorflow/tools/pip_package:prebuilt_wheel_import_api_packages_test_gpu -//tensorflow/compiler/tf2tensorrt/... -//tensorflow/core/tpu/... -//tensorflow/lite/... -//tensorflow/tools/toolchains/...
 # ARM64 WHEEL
 test:linux_arm64_wheel_test_filters --test_tag_filters=-no_oss,-tf_tosa,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310,-no_oss_py313
 test:linux_arm64_wheel_test_filters --build_tag_filters=-no_oss,-tf_tosa,-no_aarch64,-oss_excluded,-oss_serial,-gpu,-tpu,-benchmark-test,-v1only,-no_oss_py38,-no_oss_py39,-no_oss_py310,-no_oss_py313

.bazelversion

@@ -1,2 +1,2 @@
-7.4.1
+7.7.0
 # NOTE: Update Bazel version in tensorflow/tools/ci_build/release/common.sh.oss

.github/workflows/osv-scanner-scheduled.yml

@@ -28,7 +28,7 @@ permissions:
 jobs:
   scan-scheduled:
     if: github.repository == 'tensorflow/tensorflow'
-    uses: "google/osv-scanner-action/.github/workflows/osv-scanner-reusable.yml@v2.2.3"
+    uses: "google/osv-scanner-action/.github/workflows/osv-scanner-reusable.yml@v2.2.4"
     with:
       scan-args: |-
         --lockfile=requirements.txt:./requirements_lock_3_9.txt

.github/workflows/scorecards-analysis.yml

@@ -55,7 +55,7 @@ jobs:
       # Upload the results as artifacts (optional). Commenting out will disable uploads of run results in SARIF
       # format to the repository Actions tab.
       - name: "Upload artifact"
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        uses: actions/upload-artifact@330a01c490aca151604b8cf639adc76d48f6c5d4 # v5.0.0
         with:
           name: SARIF file
           path: results.sarif
@@ -64,6 +64,6 @@ jobs:
       # Upload the results to GitHub's code scanning dashboard (optional).
       # Commenting out will disable upload of results to your repo's Code Scanning dashboard
       - name: "Upload to code-scanning"
-        uses: github/codeql-action/upload-sarif@3599b3baa15b485a2e49ef411a7a4bb2452e7f93 # v3.29.5
+        uses: github/codeql-action/upload-sarif@0499de31b99561a6d14a36a5f662c2a54f91beee # v3.29.5
         with:
           sarif_file: results.sarif

.github/workflows/stale-issues.yml

@@ -31,7 +31,7 @@ jobs:
       pull-requests: write
     steps:
       - name: Awaiting response issues
-        uses: actions/stale@3a9db7e6a41a89f618792c92c0e97cc736e1b13f # v10.0.0
+        uses: actions/stale@5f858e3efba33a5ca4407a664cc011ad407f2008 # v10.1.0
         with:
           #Comma separated list of labels that can be assigned to issues to exclude them from being marked as stale
           exempt-issue-labels: 'override-stale'
@@ -59,7 +59,7 @@ jobs:
           close-pr-message: "This PR was closed because it has been inactive for 14 days since being marked as stale. Please reopen if you'd like to work on this further."
           repo-token: ${{ secrets.GITHUB_TOKEN }}
       - name: Contribution issues
-        uses: actions/stale@3a9db7e6a41a89f618792c92c0e97cc736e1b13f # v10.0.0
+        uses: actions/stale@5f858e3efba33a5ca4407a664cc011ad407f2008 # v10.1.0
         with:
           #Comma separated list of labels that can be assigned to issues to exclude them from being marked as stale
           exempt-issue-labels: 'override-stale'

RELEASE.md

@@ -22,6 +22,10 @@
 * `tf.lite`
     * Adds int8 and int16x8 support for SQRT operator.
     * Adds int16x8 support for EQUAL and NOT_EQUAL operators.
+    * Adds support for int2 type.
+    * Adds support for int2/int4 in tfl.cast .
+    * Adds support for SRQ int2 in tfl.fully_connected.
+    * Adds support for int4 in tfl.slice.
 
 ### Bug Fixes and Other Changes
 

ci/official/containers/ml_build/Dockerfile

@@ -12,14 +12,6 @@ COPY builder.packages.txt /builder.packages.txt
 
 RUN /setup.sources.sh && /setup.packages.sh /builder.packages.txt
 
-# Install devtoolset-9 in /dt9 with glibc 2.17 and libstdc++ 4.8, for building
-# manylinux2014-compatible packages.
-COPY builder.devtoolset/fixlinks.sh /fixlinks.sh
-COPY builder.devtoolset/rpm-patch.sh /rpm-patch.sh
-COPY builder.devtoolset/build_devtoolset.sh /build_devtoolset.sh
-COPY builder.devtoolset/glibc2.17-inline.patch /glibc2.17-inline.patch
-RUN /build_devtoolset.sh devtoolset-9 /dt9
-
 # Setup Python
 COPY setup.python.sh /setup.python.sh
 COPY builder.requirements.txt /builder.requirements.txt
@@ -56,9 +48,6 @@ RUN ln -sf /usr/bin/python3.12 /usr/bin/python3
 RUN ln -sf /usr/bin/python3.12 /usr/bin/python
 RUN ln -sf /usr/lib/python3.12 /usr/lib/tf_python
 
-# Make sure clang is on the path
-RUN ln -s /usr/lib/llvm-18/bin/clang /usr/bin/clang
-
 # Link the compat driver to the location if available.
 RUN if [ -e "/usr/local/cuda/compat/libcuda.so.1" ]; then ln -s /usr/local/cuda/compat/libcuda.so.1 /usr/lib/x86_64-linux-gnu/libcuda.so.1; fi
 

ci/official/containers/ml_build/builder.packages.txt

@@ -1,28 +1,9 @@
-# Packages to be installed for the new Docker image.
-
-# Packages needed to build devtoolset
-file
-flex
-g++
-make
-patch
-rpm2cpio
-unar
-wget
-xz-utils
-cpio
-
 # Other build-related tools
 apt-transport-https
 autoconf
 automake
 build-essential
 ca-certificates
-llvm-18
-clang-18
-clang-tidy-18
-lld-18
-clang-format-12
 curl
 git
 parallel
@@ -32,4 +13,6 @@ unzip
 zip
 openjdk-21-jdk
 vim
+wget
 jq
+file

ci/official/containers/ml_build/builder.requirements.txt

@@ -5,6 +5,9 @@ id
 urllib3
 requests
 
+# For XLA
+pyyaml
+
 # For JAX
 build ~= 1.2.2
 # uv is faster than pip for installing Python packages.

ci/official/containers/ml_build/cuda13.0_cudnn9.15.packages.txt

@@ -0,0 +1,23 @@
+# All required CUDA packages
+cuda-compat-13-0
+cuda-command-line-tools-13-0
+cuda-cudart-dev-13-0
+cuda-nvcc-13-0
+cuda-cupti-13-0
+cuda-nvprune-13-0
+cuda-libraries-13-0
+cuda-libraries-dev-13-0
+cuda-nvml-dev-13-0
+libcufft-13-0
+libcurand-13-0
+libcusolver-dev-13-0
+libcusparse-dev-13-0
+libcublas-13-0
+libcublas-dev-13-0
+libnccl-dev=2.27.7-1+cuda13.0
+libnccl2=2.27.7-1+cuda13.0
+# CuDNN: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html#ubuntu-network-installation
+libcudnn9-headers-cuda-13=9.15.1.9-1
+libcudnn9-static-cuda-13=9.15.1.9-1
+libcudnn9-dev-cuda-13=9.15.1.9-1
+libcudnn9-cuda-13=9.15.1.9-1

ci/official/containers/ml_build/setup.python.sh

@@ -45,16 +45,6 @@ fi
 
 /setup.packages.sh pythons.txt
 
-# Re-link pyconfig.h from x86_64-linux-gnu into the devtoolset directory
-# for any Python version present
-pushd /usr/include/x86_64-linux-gnu
-for f in $(ls | grep python); do
-  # set up symlink for devtoolset-9
-  rm -f /dt9/usr/include/x86_64-linux-gnu/$f
-  ln -s /usr/include/x86_64-linux-gnu/$f /dt9/usr/include/x86_64-linux-gnu/$f
-done
-popd
-
 # Python 3.10 include headers fix:
 # sysconfig.get_path('include') incorrectly points to /usr/local/include/python
 # map /usr/include/python3.10 to /usr/local/include/python3.10

ci/official/envs/linux_arm64

@@ -28,5 +28,5 @@ TFCI_OUTPUT_DIR=build_output
 TFCI_WHL_AUDIT_ENABLE=1
 TFCI_WHL_AUDIT_PLAT=manylinux2014_aarch64
 TFCI_WHL_BAZEL_TEST_ENABLE=1
-TFCI_WHL_SIZE_LIMIT=265M
+TFCI_WHL_SIZE_LIMIT=270M
 TFCI_WHL_SIZE_LIMIT_ENABLE=1

ci/official/envs/windows_x86_2022

@@ -15,7 +15,7 @@
 TFCI_DOCKER_ENABLE=1
 TFCI_DOCKER_PULL_ENABLE=1
 TFCI_DOCKER_IMAGE="gcr.io/tensorflow-testing/tf-win2022@sha256:915cb093630432c38b028f56bd31116a5559ebbc688d427b6092d86828ae03bc"
-TFCI_BAZEL_BAZELRC_ARGS="--output_user_root=C:/t"
+TFCI_BAZEL_BAZELRC_ARGS="--output_user_root=C:/x"
 TFCI_BAZEL_COMMON_ARGS="--repo_env=HERMETIC_PYTHON_VERSION=$TFCI_PYTHON_VERSION --repo_env=USE_PYWRAP_RULES=True --config=windows_x86_cpu_2022"
 TFCI_BAZEL_TARGET_SELECTING_CONFIG_PREFIX=windows_x86_cpu_2022
 TFCI_BUILD_PIP_PACKAGE_WHEEL_NAME_ARG="--repo_env=WHEEL_NAME=tensorflow"

ci/official/requirements_updater/numpy1_requirements/requirements.in

@@ -1,7 +1,7 @@
 # Requirements for NumPy 1.x
 numpy ~= 1.26.0
 wheel ~= 0.41.2
-h5py >= 3.11.0
+h5py >= 3.11.0, < 3.15.0
 lit ~= 17.0.2
 opt_einsum == 3.3.0
 astunparse == 1.6.3

ci/official/requirements_updater/requirements.in

@@ -1,7 +1,7 @@
 # Note that numpy 2.1.0 does not support python 3.9
 numpy >= 2.0.0, < 2.2.0
 wheel ~= 0.41.2
-h5py >= 3.11.0
+h5py >= 3.11.0, < 3.15.0
 lit ~= 17.0.2
 opt_einsum == 3.3.0
 astunparse == 1.6.3

ci/official/utilities/setup_docker.sh

@@ -62,6 +62,12 @@ if ! docker container inspect tf >/dev/null 2>&1 ; then
     # Additional setup is contained in ci/official/envs/rbe.
     CONTAINER_IP_ADDR=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' tf)
     netsh advfirewall firewall add rule name="Allow Metadata Proxy" dir=in action=allow protocol=TCP localport=80 remoteip="$CONTAINER_IP_ADDR"
+
+    # Stop non-essential indexing and link tracking services that
+    # may lock new files or symlinks.
+    # They may be causing sporadic "Permission denied" errors during Bazel builds.
+    # b/461500885
+    docker exec tf powershell -NoProfile -Command 'Stop-Service -Name SysMain,DiagTrack -Force -ErrorAction SilentlyContinue'
   fi
 
 fi

tensorflow/BUILD

@@ -1033,6 +1033,7 @@ package_group(
         "//tensorflow_models/google/recml/...",
         "//third_party/cloud_tpu/convergence_tools/sdc_monitoring/...",
         "//third_party/cloud_tpu/inference_converter/...",
+        "//third_party/pathways/...",
         "//third_party/py/cloud_ml_autoflow/...",
         "//third_party/py/envlogger/...",
         "//third_party/py/gldm/...",
@@ -1180,38 +1181,31 @@ tf_cc_shared_library(
     linkstatic = 1,
     per_os_targets = True,
     roots = [
-                "//tensorflow/c/experimental/filesystem:filesystem_interface",
-                "//tensorflow/c/experimental/stream_executor:stream_executor",
-                "//tensorflow/c:env",
-                "//tensorflow/c:kernels",
-                "//tensorflow/c:kernels_experimental",
-                "//tensorflow/c:logging",
-                "//tensorflow/c:ops",
-                "//tensorflow/cc/saved_model:fingerprinting_impl",
-                "//tensorflow/cc/saved_model:loader_lite_impl",
-                "//tensorflow/cc/saved_model:metrics_impl",
-                "//tensorflow/compiler/tf2tensorrt:op_converter_registry_impl",
-                "//tensorflow/core/common_runtime:core_cpu_impl",
-                "//tensorflow/core/common_runtime/gpu:gpu_runtime_impl",
-                "//tensorflow/core/common_runtime/pluggable_device:pluggable_device_runtime_impl",
-                "//tensorflow/core:framework_internal_impl",
-                "//tensorflow/core/framework:tensor",
-                "//tensorflow/core/grappler/optimizers:custom_graph_optimizer_registry_impl",
-                "//tensorflow/core:lib_internal_impl",
-                "//tensorflow/core/profiler:profiler_impl",
-                "//tensorflow/core/util:determinism",  # Must be linked and exported to libtensorflow_framework.so.
-                "//tensorflow/lite/kernels/shim:tf_kernel_shim",
-                "@local_xla//xla/stream_executor:stream_executor_impl",
-                "@local_xla//xla/tsl/framework:bfc_allocator",
-                "@local_xla//xla/tsl/framework:metrics",
-            ] + tf_additional_binary_deps() +
-            # TODO(b/259305727): Remove this select and include captured_function in macos builds.
-            select({
-                "//tensorflow:macos": [],
-                "//conditions:default": [
-                    "//tensorflow/core/data:captured_function",
-                ],
-            }),
+        "//tensorflow/c/experimental/filesystem:filesystem_interface",
+        "//tensorflow/c/experimental/stream_executor:stream_executor",
+        "//tensorflow/c:env",
+        "//tensorflow/c:kernels",
+        "//tensorflow/c:kernels_experimental",
+        "//tensorflow/c:ops",
+        "//tensorflow/cc/saved_model:fingerprinting_impl",
+        "//tensorflow/cc/saved_model:loader_lite_impl",
+        "//tensorflow/cc/saved_model:metrics_impl",
+        "//tensorflow/compiler/tf2tensorrt:op_converter_registry_impl",
+        "//tensorflow/core/common_runtime:core_cpu_impl",
+        "//tensorflow/core/common_runtime/gpu:gpu_runtime_impl",
+        "//tensorflow/core/common_runtime/pluggable_device:pluggable_device_runtime_impl",
+        "//tensorflow/core:framework_internal_impl",
+        "//tensorflow/core/framework:tensor",
+        "//tensorflow/core/grappler/optimizers:custom_graph_optimizer_registry_impl",
+        "//tensorflow/core:lib_internal_impl",
+        "//tensorflow/core/profiler:profiler_impl",
+        "//tensorflow/core/util:determinism",  # Must be linked and exported to libtensorflow_framework.so.
+        "//tensorflow/lite/kernels/shim:tf_kernel_shim",
+        "@local_xla//xla/stream_executor:stream_executor_impl",
+        "@local_xla//xla/tsl/framework:bfc_allocator",
+        "@local_xla//xla/tsl/framework:metrics",
+        "//tensorflow/core/data:captured_function",
+    ] + tf_additional_binary_deps(),
     soversion = VERSION,
     static_deps = PACKAGE_STATIC_DEPS,
     visibility = ["//visibility:public"],

tensorflow/c/BUILD

@@ -298,7 +298,6 @@ tf_cuda_library(
         ],
         "//conditions:default": [
             ":env",
-            ":logging",
             ":tf_status",
             ":tf_tensor",
             "//tensorflow/c/experimental/filesystem:modular_filesystem",
@@ -325,18 +324,6 @@ tf_cuda_library(
     alwayslink = 1,
 )
 
-cc_library(
-    name = "logging",
-    srcs = ["logging.cc"],
-    hdrs = ["logging.h"],
-    visibility = ["//visibility:public"],
-    deps = [
-        ":c_api_macros",
-        "//tensorflow/core/platform:logging",
-        "//tensorflow/core/platform:stringprintf",
-    ],
-)
-
 tf_cuda_library(
     name = "tf_status_internal",
     hdrs = [

tensorflow/c/c_api_function_test.cc

@@ -1171,7 +1171,7 @@ TEST_F(CApiFunctionTest, InvalidOutputTensor_BadNodePtr) {
   EXPECT_EQ(TF_INVALID_ARGUMENT, TF_GetCode(s_));
   EXPECT_EQ(string("Node is null\n\tEncountered while processing output 0 "
                    "from function 'MyFunc'"),
-            string(TF_Message(s_)));
+            std::string(TF_Message(s_)));
 }
 
 TEST_F(CApiFunctionTest, NodeMissingInput) {

tensorflow/c/c_api_test.cc

@@ -2478,7 +2478,7 @@ TEST_F(CApiAttributesTest, Names) {
 
   TF_OperationGetAttrName(oper, 0, value.get(), s_);
   EXPECT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
-  EXPECT_EQ("v", string(static_cast<const char*>(value.get()), 1));
+  EXPECT_EQ("v", std::string(static_cast<const char*>(value.get()), 1));
 }
 
 TEST_F(CApiAttributesTest, Errors) {