[AutoGluon] Add AutoGluon v1.1.0 (#3867)

* Draft AutoGluon DLC 1.1.0 * update tests * capping mmcv * vulnerability scans * Revert "vulnerability scans" This reverts commit 3f2a9de. * ray upgrade * fix vulnerabilities * update * update training * updating inference * update packages * add setuptools * test * fix tqdm vulnerability * cleanup * cleanup * revert toml * fix toml * test version * revert * revert toml * revert toml * update available images * scan for vulnerabilities * fix ray vulnerabilities * fix vulnerability for gpu images * revert toml --------- Co-authored-by: Ubuntu <[email protected]>
aws · May 9, 2024 · db83f12 · db83f12
1 parent bc13cd9
commit db83f12
Show file tree

Hide file tree

Showing 18 changed files with 1,266 additions and 14 deletions.
diff --git a/autogluon/inference/buildspec-1-0-0.yml b/autogluon/inference/buildspec-1-0-0.yml
@@ -0,0 +1,53 @@
+account_id: &ACCOUNT_ID <set-$ACCOUNT_ID-in-environment>
+region: &REGION <set-$REGION-in-environment>
+framework: &FRAMEWORK autogluon
+version: &VERSION 1.0.0
+short_version: &SHORT_VERSION 1.0
+arch_type: x86
+
+repository_info:
+  inference_repository: &INFERENCE_REPOSITORY
+    image_type: &INFERENCE_IMAGE_TYPE inference
+    root: !join [ *FRAMEWORK, "/", *INFERENCE_IMAGE_TYPE ]
+    repository_name: &REPOSITORY_NAME !join [pr, "-", *FRAMEWORK, "-", *INFERENCE_IMAGE_TYPE]
+    repository: &REPOSITORY !join [ *ACCOUNT_ID, .dkr.ecr., *REGION, .amazonaws.com/, *REPOSITORY_NAME ]
+
+context:
+  inference_context: &INFERENCE_CONTEXT
+    torchserve-entrypoint:
+      source: ../build_artifacts/inference/torchserve-entrypoint.py
+      target: torchserve-entrypoint.py
+    config:
+      source: ../build_artifacts/inference/config.properties
+      target: config.properties
+    deep_learning_container:
+      source: ../../src/deep_learning_container.py
+      target: deep_learning_container.py
+
+images:
+  BuildAutogluonCPUInferencePy3DockerImage:
+    <<: *INFERENCE_REPOSITORY
+    build: &AUTOGLUON_CPU_INFERENCE_PY3 false
+    image_size_baseline: 6399
+    device_type: &DEVICE_TYPE cpu
+    python_version: &DOCKER_PYTHON_VERSION py3
+    tag_python_version: &TAG_PYTHON_VERSION py310
+    os_version: &OS_VERSION ubuntu20.04
+    tag: !join [ *VERSION, "-", *DEVICE_TYPE, "-", *TAG_PYTHON_VERSION, "-", *OS_VERSION ]
+    docker_file: !join [ docker/, *SHORT_VERSION, /, *DOCKER_PYTHON_VERSION, /Dockerfile., *DEVICE_TYPE ]
+    context:
+      <<: *INFERENCE_CONTEXT
+
+  BuildAutogluonGPUInferencePy3DockerImage:
+    <<: *INFERENCE_REPOSITORY
+    build: &AUTOGLUON_GPU_INFERENCE_PY3 false
+    image_size_baseline: 16613
+    device_type: &DEVICE_TYPE gpu
+    python_version: &DOCKER_PYTHON_VERSION py3
+    tag_python_version: &TAG_PYTHON_VERSION py310
+    cuda_version: &CUDA_VERSION cu118
+    os_version: &OS_VERSION ubuntu20.04
+    tag: !join [ *VERSION, "-", *DEVICE_TYPE, "-", *TAG_PYTHON_VERSION, "-", *CUDA_VERSION, "-", *OS_VERSION ]
+    docker_file: !join [ docker/, *SHORT_VERSION, /, *DOCKER_PYTHON_VERSION, /, *CUDA_VERSION, /Dockerfile., *DEVICE_TYPE ]
+    context:
+      <<: *INFERENCE_CONTEXT
diff --git a/autogluon/inference/buildspec.yml b/autogluon/inference/buildspec.yml
@@ -1,8 +1,8 @@
 account_id: &ACCOUNT_ID <set-$ACCOUNT_ID-in-environment>
 region: &REGION <set-$REGION-in-environment>
 framework: &FRAMEWORK autogluon
-version: &VERSION 1.0.0
-short_version: &SHORT_VERSION 1.0
+version: &VERSION 1.1.0
+short_version: &SHORT_VERSION 1.1
 arch_type: x86
 
 repository_info:
@@ -41,7 +41,7 @@ images:
   BuildAutogluonGPUInferencePy3DockerImage:
     <<: *INFERENCE_REPOSITORY
     build: &AUTOGLUON_GPU_INFERENCE_PY3 false
-    image_size_baseline: 16613
+    image_size_baseline: 19456
     device_type: &DEVICE_TYPE gpu
     python_version: &DOCKER_PYTHON_VERSION py3
     tag_python_version: &TAG_PYTHON_VERSION py310

diff --git a/autogluon/inference/docker/1.1/py3/Dockerfile.cpu b/autogluon/inference/docker/1.1/py3/Dockerfile.cpu
@@ -0,0 +1,75 @@
+ARG PYTHON_VERSION=3.10.8
+
+FROM 763104351884.dkr.ecr.us-west-2.amazonaws.com/pytorch-inference:2.1.0-cpu-py310-ubuntu20.04-sagemaker
+
+# Specify accept-bind-to-port LABEL for inference pipelines to use SAGEMAKER_BIND_TO_PORT
+# https://docs.aws.amazon.com/sagemaker/latest/dg/inference-pipeline-real-time.html
+LABEL com.amazonaws.sagemaker.capabilities.accept-bind-to-port=true
+# Specify multi-models LABEL to indicate container is capable of loading and serving multiple models concurrently
+# https://docs.aws.amazon.com/sagemaker/latest/dg/build-multi-model-build-container.html
+LABEL com.amazonaws.sagemaker.capabilities.multi-models=true
+
+LABEL maintainer="Amazon AI"
+LABEL dlc_major_version="1"
+
+RUN apt-get update \
+ && apt-get -y upgrade \
+ && apt-get autoremove -y \
+ && apt-get install tesseract-ocr -y \
+ && apt-get clean \
+ && rm -rf /var/lib/apt/lists/*
+
+ARG AUTOGLUON_VERSION=1.1.0
+
+# Upgrading pip and installing/updating Python dependencies
+# Comments are added to explain the reason behind each update
+RUN pip install --no-cache-dir -U --trusted-host pypi.org --trusted-host files.pythonhosted.org pip \
+ && pip install --no-cache-dir -U wheel \
+ && pip uninstall -y dataclasses \
+ && pip install --no-cache-dir -U numpy numba \
+# Install AutoGluon, ensuring no vulnerable dependencies are left behind
+ && pip install --no-cache-dir -U autogluon==${AUTOGLUON_VERSION} \
+# Update requests to a secure version that fixes CVE-2023-32681
+ && pip install --no-cache-dir -U requests>=2.31.0 \  
+ && mim install -q mmcv==2.1.0 --timeout 600 \
+ && pip install --no-cache-dir -U mmdet>=3.0.0 \
+ && pip install --no-cache-dir -U "mmengine" \
+# Update ray to a secure version that fixes CVE-2023-48022
+ && pip install --no-cache-dir -U ray>=2.10.0 \
+# Update transformers to secure versions that fixes CVE-2023-7018
+ && pip install --no-cache-dir -U setuptools>=65.5.1 \  
+ # Update pillow to a secure version that fixes CVE-2023-50447
+ && pip install --no-cache-dir -U pillow>=10.2.0 \
+ # Update Jinja2 to a secure version that fixes vulnerability id 64227
+ && pip install --no-cache-dir -U Jinja2>=3.1.3 \
+ # Update aiohttp to a secure version that fixes vulnerability id 64642
+ && pip install --no-cache-dir -U aiohttp>=3.9.2 \
+ # below adjustments are done to be consistent with the training container
+ && pip install --no-cache-dir -U scikit-learn==1.4.0 \
+ && pip install --no-cache-dir -U scipy==1.12.0 \
+ && pip install --no-cache-dir -U joblib==1.4.2 \
+ # update tqdm to a secure version that fixes CVE-2024-34062
+ && pip install --no-cache-dir -U tqdm>=4.66.3
+
+
+
+# add TS entrypoint
+COPY config.properties /home/model-server
+
+COPY torchserve-entrypoint.py /usr/local/bin/dockerd-entrypoint.py
+RUN chmod +x /usr/local/bin/dockerd-entrypoint.py
+
+RUN HOME_DIR=/root \
+ && curl -o ${HOME_DIR}/oss_compliance.zip https://aws-dlinfra-utilities.s3.amazonaws.com/oss_compliance.zip \
+ && unzip -o ${HOME_DIR}/oss_compliance.zip -d ${HOME_DIR}/ \
+ && cp ${HOME_DIR}/oss_compliance/test/testOSSCompliance /usr/local/bin/testOSSCompliance \
+ && chmod +x /usr/local/bin/testOSSCompliance \
+ && chmod +x ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh \
+ && ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh ${HOME_DIR} python \
+ && rm -rf ${HOME_DIR}/oss_compliance*
+
+RUN curl -o /licenses-autogluon.txt https://autogluon.s3.us-west-2.amazonaws.com/licenses/THIRD-PARTY-LICENSES.txt
+
+EXPOSE 8080 8081
+ENTRYPOINT ["python", "/usr/local/bin/dockerd-entrypoint.py"]
+CMD ["torchserve", "--start", "--ts-config", "/home/model-server/config.properties", "--model-store", "/home/model-server/"]
diff --git a/autogluon/inference/docker/1.1/py3/Dockerfile.cpu.os_scan_allowlist.json b/autogluon/inference/docker/1.1/py3/Dockerfile.cpu.os_scan_allowlist.json
@@ -0,0 +1,209 @@
+{
+    "linux": [
+        {
+            "description": " An out-of-bounds read vulnerability was found in smbCalcSize in fs/smb/client/netmisc.c in the Linux Kernel. This issue could allow a local attacker to crash the system or leak internal kernel information.",
+            "vulnerability_id": "CVE-2023-6606",
+            "name": "CVE-2023-6606",
+            "package_name": "linux",
+            "package_details": {
+                "file_path": null,
+                "name": "linux",
+                "package_manager": "OS",
+                "version": "5.4.0",
+                "release": "169.187"
+            },
+            "remediation": {
+                "recommendation": {
+                    "text": "None Provided"
+                }
+            },
+            "cvss_v3_score": 7.1,
+            "cvss_v30_score": 0.0,
+            "cvss_v31_score": 7.1,
+            "cvss_v2_score": 0.0,
+            "cvss_v3_severity": "HIGH",
+            "source_url": "https://people.canonical.com/~ubuntu-security/cve/2023/CVE-2023-6606.html",
+            "source": "UBUNTU_CVE",
+            "severity": "HIGH",
+            "status": "ACTIVE",
+            "title": "CVE-2023-6606 - linux",
+            "reason_to_ignore": "vulnerability from Linux kernel 5.4.0 in base image"
+        },
+        {
+            "description": "\n It was discovered that the CIFS network file system implementation in\n the Linux kernel did not properly validate the server frame size in\n certain situation, leading to an out-of-bounds read vulnerability. An\n attacker could use this to construct a malicious CIFS image that, when\n operated on, could cause a denial of service (system crash) or possibly\n expose sensitive information.",
+            "vulnerability_id": "CVE-2023-6606",
+            "name": "CVE-2023-6606",
+            "package_name": "linux",
+            "package_details": {
+                "file_path": null,
+                "name": "linux",
+                "package_manager": "OS",
+                "version": "5.4.0",
+                "release": "170.188"
+            },
+            "remediation": {
+                "recommendation": {
+                    "text": "None Provided"
+                }
+            },
+            "cvss_v3_score": 7.1,
+            "cvss_v30_score": 0.0,
+            "cvss_v31_score": 7.1,
+            "cvss_v2_score": 0.0,
+            "cvss_v3_severity": "HIGH",
+            "source_url": "https://people.canonical.com/~ubuntu-security/cve/2023/CVE-2023-6606.html",
+            "source": "UBUNTU_CVE",
+            "severity": "HIGH",
+            "status": "ACTIVE",
+            "title": "CVE-2023-6606 - linux",
+            "reason_to_ignore": "vulnerability from Linux kernel 5.4.0 in base image"
+        },
+        {
+            "description": "\n It was discovered that a race condition existed in the Rose X.25\n protocol implementation in the Linux kernel, leading to a use-after-\n free vulnerability. A local attacker could use this to cause a denial\n of service (system crash) or possibly execute arbitrary code.",
+            "vulnerability_id": "CVE-2023-51782",
+            "name": "CVE-2023-51782",
+            "package_name": "linux",
+            "package_details": {
+              "file_path": null,
+              "name": "linux",
+              "package_manager": "OS",
+              "version": "5.4.0",
+              "release": "171.189"
+            },
+            "remediation": {
+              "recommendation": {
+                "text": "None Provided"
+              }
+            },
+            "cvss_v3_score": 7.0,
+            "cvss_v30_score": 0.0,
+            "cvss_v31_score": 7.0,
+            "cvss_v2_score": 0.0,
+            "cvss_v3_severity": "HIGH",
+            "source_url": "https://people.canonical.com/~ubuntu-security/cve/2023/CVE-2023-51782.html",
+            "source": "UBUNTU_CVE",
+            "severity": "HIGH",
+            "status": "ACTIVE",
+            "title": "CVE-2023-51782 - linux",
+            "reason_to_ignore": "fix not available yet"
+          },
+          {
+            "description": "\n It was discovered that a race condition existed in the ATM\n (Asynchronous Transfer Mode) subsystem of the Linux kernel, leading\n to a use-after-free vulnerability. A local attacker could use this\n to cause a denial of service (system crash) or possibly execute\n arbitrary code.",
+            "vulnerability_id": "CVE-2023-51780",
+            "name": "CVE-2023-51780",
+            "package_name": "linux",
+            "package_details": {
+              "file_path": null,
+              "name": "linux",
+              "package_manager": "OS",
+              "version": "5.4.0",
+              "release": "171.189"
+            },
+            "remediation": {
+              "recommendation": {
+                "text": "None Provided"
+              }
+            },
+            "cvss_v3_score": 7.0,
+            "cvss_v30_score": 0.0,
+            "cvss_v31_score": 7.0,
+            "cvss_v2_score": 0.0,
+            "cvss_v3_severity": "HIGH",
+            "source_url": "https://people.canonical.com/~ubuntu-security/cve/2023/CVE-2023-51780.html",
+            "source": "UBUNTU_CVE",
+            "severity": "HIGH",
+            "status": "ACTIVE",
+            "title": "CVE-2023-51780 - linux",
+            "reason_to_ignore": "fix not available yet"
+          },
+          {
+            "description": "\n It was discovered that a race condition existed in the AppleTalk\n networking subsystem of the Linux kernel, leading to a use-after-free\n vulnerability. A local attacker could use this to cause a denial of\n service (system crash) or possibly execute arbitrary code.",
+            "vulnerability_id": "CVE-2023-51781",
+            "name": "CVE-2023-51781",
+            "package_name": "linux",
+            "package_details": {
+              "file_path": null,
+              "name": "linux",
+              "package_manager": "OS",
+              "version": "5.4.0",
+              "release": "171.189"
+            },
+            "remediation": {
+              "recommendation": {
+                "text": "None Provided"
+              }
+            },
+            "cvss_v3_score": 7.0,
+            "cvss_v30_score": 0.0,
+            "cvss_v31_score": 7.0,
+            "cvss_v2_score": 0.0,
+            "cvss_v3_severity": "HIGH",
+            "source_url": "https://people.canonical.com/~ubuntu-security/cve/2023/CVE-2023-51781.html",
+            "source": "UBUNTU_CVE",
+            "severity": "HIGH",
+            "status": "ACTIVE",
+            "title": "CVE-2023-51781 - linux",
+            "reason_to_ignore": "fix not available yet"
+          }
+    ],
+    "ray": [
+        {
+            "description": "Anyscale Ray 2.6.3 and 2.8.0 allows a remote attacker to execute arbitrary code via the job submission API. NOTE: the vendor's position is that this report is irrelevant because Ray, as stated in its documentation, is not intended for use outside of a strictly controlled network environment",
+            "vulnerability_id": "CVE-2023-48022",
+            "name": "CVE-2023-48022",
+            "package_name": "ray",
+            "package_details": {
+              "file_path": "opt/conda/lib/python3.10/site-packages/ray-2.20.0.dist-info/METADATA",
+              "name": "ray",
+              "package_manager": "PYTHONPKG",
+              "version": "2.20.0",
+              "release": null
+            },
+            "remediation": {
+              "recommendation": {
+                "text": "None Provided"
+              }
+            },
+            "cvss_v3_score": 9.8,
+            "cvss_v30_score": 0,
+            "cvss_v31_score": 9.8,
+            "cvss_v2_score": 0,
+            "cvss_v3_severity": "CRITICAL",
+            "source_url": "https://nvd.nist.gov/vuln/detail/CVE-2023-48022",
+            "source": "NVD",
+            "severity": "CRITICAL",
+            "status": "ACTIVE",
+            "title": "CVE-2023-48022 - ray",
+            "reason_to_ignore": "fix not available yet"
+        },
+        {
+            "description": "Anyscale Ray 2.6.3 and 2.8.0 allows a remote attacker to execute arbitrary code via the job submission API. NOTE: the vendor's position is that this report is irrelevant because Ray, as stated in its documentation, is not intended for use outside of a strictly controlled network environment",
+            "vulnerability_id": "CVE-2023-48022",
+            "name": "CVE-2023-48022",
+            "package_name": "ray",
+            "package_details": {
+              "file_path": "opt/conda/lib/python3.10/site-packages/ray-2.21.0.dist-info/METADATA",
+              "name": "ray",
+              "package_manager": "PYTHONPKG",
+              "version": "2.21.0",
+              "release": null
+            },
+            "remediation": {
+              "recommendation": {
+                "text": "None Provided"
+              }
+            },
+            "cvss_v3_score": 9.8,
+            "cvss_v30_score": 0,
+            "cvss_v31_score": 9.8,
+            "cvss_v2_score": 0,
+            "cvss_v3_severity": "CRITICAL",
+            "source_url": "https://nvd.nist.gov/vuln/detail/CVE-2023-48022",
+            "source": "NVD",
+            "severity": "CRITICAL",
+            "status": "ACTIVE",
+            "title": "CVE-2023-48022 - ray",
+            "reason_to_ignore": "fix not available yet"
+        }
+    ]
+}
diff --git a/autogluon/inference/docker/1.1/py3/Dockerfile.cpu.py_scan_allowlist.json b/autogluon/inference/docker/1.1/py3/Dockerfile.cpu.py_scan_allowlist.json
@@ -0,0 +1,7 @@
+{
+    "50916": "Pydantic 1.10.2 prevents long strings as int inputs to fix CVE-2020-10735 - upstream dependencies are still not patched",
+    "51358": "Safety 2.2.0 updates its dependency 'dparse' to include a security fix. - not packaged with container, result of security scanning process",
+    "65213": "PyOpenSSL doesn't have this fix yet - the issue only applicable to PowerPC architecture and not applicable to this container",
+    "67599": "pip - No fix for this yet",
+    "65345": "torchserve - we do not use torchserve, this vulnerability leaks from an upstream base image"
+}