Neuron SDK Release 2.20.2

ivashkst · jeffhataws · natemail-aws · ivashkst · commit a574acbad93f · 2024-11-21T03:04:37.000Z
Release notes for Neuron SDK Release 2.20.2

---------

Co-authored-by: Jeffrey Huynh &lt;jthuynh@amazon.com&gt;
Co-authored-by: Nathan Mailhot &lt;natemail@amazon.com&gt;
Co-authored-by: Ahsan Khan &lt;amzahsa@amazon.com&gt;
Co-authored-by: Mounik Chinthapanti &lt;mounchin@amazon.com&gt;
Co-authored-by: Nathan Mailhot &lt;120211561+natemail-aws@users.noreply.github.com&gt;
Co-authored-by: Ryan King &lt;kingrj@6amazon.com&gt;
Co-authored-by: musunita &lt;62577111+musunita@users.noreply.github.com&gt;
Co-authored-by: Vikas Paliwal &lt;paliwavi@amazon.com&gt;
Co-authored-by: Pradeep Roy &lt;lroyprad@amazon.com&gt;
Co-authored-by: Esha Lakhotia &lt;72316944+eshalakhotia@users.noreply.github.com&gt;
Co-authored-by: Nicholas Waldron &lt;waldronn@amazon.com&gt;
Co-authored-by: Roopnath &lt;122940636+rgrandhiamzn@users.noreply.github.com&gt;
diff --git a/conf.py b/conf.py
@@ -157,7 +157,7 @@
 
 #top_banner_message="<span>&#9888;</span><a class='reference internal' style='color:white;' href='https://awsdocs-neuron.readthedocs-hosted.com/en/latest/general/setup/setup-troubleshooting.html#gpg-key-update'>  Neuron repository GPG key for Ubuntu installation has expired, see instructions how to update! </a>"
 
-top_banner_message="Neuron 2.20.1 is released! check <a class='reference internal' style='color:white;' href='https://awsdocs-neuron.readthedocs-hosted.com/en/latest/release-notes/index.html#latest-neuron-release'> What's New  </a> and <a class='reference internal' style='color:white;' href='https://awsdocs-neuron.readthedocs-hosted.com/en/latest/general/announcements/index.html'> Announcements  </a>"
+top_banner_message="Neuron 2.20.2 is released! check <a class='reference internal' style='color:white;' href='https://awsdocs-neuron.readthedocs-hosted.com/en/latest/release-notes/index.html#latest-neuron-release'> What's New  </a> and <a class='reference internal' style='color:white;' href='https://awsdocs-neuron.readthedocs-hosted.com/en/latest/general/announcements/index.html'> Announcements  </a>"
 
 html_theme = "sphinx_book_theme"
 html_theme_options = {
diff --git a/release-notes/containers/neuron-dlc.rst b/release-notes/containers/neuron-dlc.rst
@@ -7,11 +7,18 @@ Neuron DLC Release Notes
    :local:
    :depth: 1
 
+Neuron 2.20.2
+-------------
+Date: 11/20/2024
+
+- Neuron 2.20.2 DLC fixes dependency bug for NxDT use case by pinning the correct torch version. 
+
 
 Neuron 2.20.1
 -------------
 
 Date: 10/25/2024
+
 - Neuron 2.20.1 DLC includes prerequisites for :ref:`nxdt_installation_guide`. Customers can expect to use NxDT out of the box.
 
 
diff --git a/release-notes/containers/neuron-k8.rst b/release-notes/containers/neuron-k8.rst
@@ -35,6 +35,16 @@ To Pull the Images from ECR:
    docker pull  public.ecr.aws/neuron/neuron-scheduler:2.x.y.z
 
 
+Neuron K8 release [2.22.20.0]
+=============================
+
+Date: 11/20/2024
+
+Bug fixes
+---------
+
+- This release addresses a stability issue in the Neuron Scheduler Extension that previously caused crashes shortly after installation.
+
 Neuron K8 release [2.22.4.0]
 ============================
 
diff --git a/release-notes/index.rst b/release-notes/index.rst
@@ -11,6 +11,23 @@ What's New
 .. _neuron-2.20.0-whatsnew:
 
 
+Neuron 2.20.2 (11/20/2024)
+---------------------------
+
+Neuron 2.20.2 release fixes a stability issue in Neuron Scheduler Extension that previously caused crashes in Kubernetes (K8) deployments. See :ref:`neuron-k8-rn`.
+
+This release also addresses a security patch update to Neuron Driver that fixes a kernel address leak issue. 
+See more on :ref:`neuron-driver-release-notes` and :ref:`neuron-runtime-rn`.
+
+Addtionally, Neuron 2.20.2 release updates ``torch-neuronx`` and ``libneuronxla`` packages to add support for ``torch-xla`` 2.1.5 package 
+which fixes checkpoint loading issues with Zero Redundancy Optimizer (ZeRO-1). See :ref:`torch-neuronx-rn` and :ref:`libneuronxla-rn`.
+
+Neuron supported DLAMIs and DLCs are updated with this release (Neuron 2.20.2 SDK). The Training DLC is also updated to address the 
+version dependency issues in NxD Training library. See :ref:`neuron-dlc-release-notes`.
+
+NxD Training library in Neuron 2.20.2 release is updated to transformers 4.36.0 package. See :ref:`neuronx-distributed-training-rn`.
+
+
 Neuron 2.20.1 (10/25/2024)
 ---------------------------
 
@@ -399,27 +416,27 @@ Release Artifacts
 Trn1 packages
 ^^^^^^^^^^^^^^
 
-.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=trn1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=trn1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.2
 
 Inf2 packages
 ^^^^^^^^^^^^^^
 
-.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf2 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf2 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.2
 
 Inf1 packages
 ^^^^^^^^^^^^^^
 
-.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.2
 
 Supported Python Versions for Inf1 packages
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
-.. program-output:: python3 src/helperscripts/n2-helper.py --list=pyversions --instance=inf1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=pyversions --instance=inf1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.2
 
 Supported Python Versions for Inf2/Trn1 packages
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
-.. program-output:: python3 src/helperscripts/n2-helper.py --list=pyversions --instance=inf2 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=pyversions --instance=inf2 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.2
 
 Supported Numpy Versions
 ^^^^^^^^^^^^^^^^^^^^^^^^
diff --git a/release-notes/libneuronxla/index.rst b/release-notes/libneuronxla/index.rst
@@ -15,6 +15,15 @@ the `PJRT <https://openxla.org/xla/pjrt_integration>`__ runtime, built using
 the `PJRT C-API plugin <https://github.com/openxla/xla/blob/5564a9220af230c6c194e37b37938fb40692cfc7/xla/pjrt/c/docs/pjrt_integration_guide.md>`__
 mechanism.
 
+Release [2.0.5347.0]
+--------------------
+Date: 11/20/2024
+
+Summary
+~~~~~~~
+
+Add support for torch-xla 2.1.5 which fixes the "list index out of range" error when using the Zero Redundancy Optimizer (ZeRO1) checkpoint loading.
+
 Release [2.0.4986.0]
 --------------------
 Date: 10/25/2024
diff --git a/release-notes/neuronx-distributed-training/neuronx-distributed-training.rst b/release-notes/neuronx-distributed-training/neuronx-distributed-training.rst
@@ -10,14 +10,25 @@ NxD Training Release Notes (``neuronx-distributed-training``)
 
 This document lists the release notes for Neuronx Distributed Training library.
 
-.. _neuronx-distributed-rn-1-0-0:
+.. _neuronx-distributed-training-rn-1-0-1:
+
+Neuronx Distributed Training [1.0.1]
+
+Date: 11/20/2024
+
+Features in this release
+------------------------
+
+* Added support for transformers 4.36.0
+
+.. _neuronx-distributed-training-rn-1-0-0:
 
 Neuronx Distributed Training [1.0.0]
 
 Date: 09/16/2024
 
-Features this release
----------------------
+Features in this release
+------------------------
 
 This is the first release of NxD Training (NxDT), NxDT is a PyTorch-based library that adds support for user-friendly distributed training experience through a YAML configuration file compatible with NeMo,, allowing users to easily set up their training workflows. At the same time, NxDT maintains flexibility, enabling users to choose between using the YAML configuration file, PyTorch Lightning Trainer, or writing their own custom training script using the NxD Core.
 The library supports PyTorch model classes including Hugging Face and Megatron-LM. Additionally, it leverages NeMo's data engineering and data science modules enabling end-to-end training workflows on NxDT, and providing a compatability with NeMo through minimal changes to the YAML configuration file for models that are already supported in NxDT. Furthermore, the functionality of the Neuron NeMo Megatron (NNM) library is now part of NxDT, ensuring a smooth migration path from NNM to NxDT.
diff --git a/release-notes/prev/content.rst b/release-notes/prev/content.rst
@@ -7,6 +7,21 @@ Previous Releases Artifacts (Neuron 2.x)
    :local:
    :depth: 1
 
+Neuron 2.20.1 (10/25/2024)
+---------------------------
+
+Trn1 packages
+^^^^^^^^^^^^^
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=trn1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+
+Inf2 packages
+^^^^^^^^^^^^^
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf2 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+
+Inf1 packages
+^^^^^^^^^^^^^
+.. program-output:: python3 src/helperscripts/n2-helper.py --list=packages --instance=inf1 --file=src/helperscripts/n2-manifest.json --neuron-version=2.20.1
+
 Neuron 2.20.0 (09/16/2024)
 ---------------------------
 
diff --git a/release-notes/runtime/aws-neuronx-dkms/index.rst b/release-notes/runtime/aws-neuronx-dkms/index.rst
@@ -15,6 +15,15 @@ Updated : 04/29/2022
 
 - In rare cases of multi-process applications running under heavy stress a model load failure my occur. This may require reloading of the Neuron Driver as a workaround.
 
+
+Neuron Driver release [2.18.20.0]
+--------------------------------
+Date: 11/20/2024
+
+Bug Fixes
+^^^^^^^^^
+* This release addresses an issue with Neuron Driver that can lead to a user-space application either gaining access to kernel addresses or providing the driver with spoofed memory handles (kernel addresses) that can be potentially used to gain elevated privileges. We would like to thank `Cossack9989 <https://github.com/Cossack9989>`_ for reporting and collaborating on this issue.
+
 Neuron Driver release [2.18.12.0]
 --------------------------------
 
diff --git a/release-notes/runtime/aws-neuronx-runtime-lib/index.rst b/release-notes/runtime/aws-neuronx-runtime-lib/index.rst
@@ -31,6 +31,14 @@ NEFF Version Runtime Version Range Notes
 2.0          >= 1.6.5.0            Starting support for 2.0 NEFFs 
 ============ ===================== ===================================
 
+Neuron Runtime Library [2.22.19.0]
+---------------------------------
+Date: 11/20/2024
+
+New in this release
+^^^^^^^^^^^^^^^^^^^
+* Minor improvements and bug fixes
+
 Neuron Runtime Library [2.22.14.0]
 ---------------------------------
 Date: 09/16/2024
diff --git a/release-notes/torch/torch-neuronx/index.rst b/release-notes/torch/torch-neuronx/index.rst
@@ -14,6 +14,16 @@ PyTorch Neuron for |Trn1|/|Inf2| is a software package that enables PyTorch
 users to train, evaluate, and perform inference on second-generation Neuron
 hardware (See: :ref:`NeuronCore-v2 <neuroncores-v2-arch>`).
 
+Release [2.1.2.2.3.2]
+----------------------
+Date: 11/20/2024
+
+Summary
+~~~~~~~
+
+This patch narrows the range of dependent libneuronxla versions to support minor version bumps
+and fixes the "list index out of range" error when using the Zero Redundancy Optimizer (ZeRO1) checkpoint loading.
+
 Release [2.1.2.2.3.1]
 ----------------------
 Date: 10/25/2024
diff --git a/src/examples/pytorch/torch-neuronx/t5-inference-tutorial.ipynb b/src/examples/pytorch/torch-neuronx/t5-inference-tutorial.ipynb
@@ -52,7 +52,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "!pip install --upgrade transformers==4.31.0 optimum-neuron==0.0.8"
+    "!pip install --upgrade transformers==4.31.0 optimum-neuron==0.0.8 sentencepiece"
    ]
   },
   {
diff --git a/src/helperscripts/n2-manifest.json b/src/helperscripts/n2-manifest.json
@@ -4,13 +4,13 @@
       {"repo_type":"rpm", "repo_url":"https://yum.repos.neuron.amazonaws.com/"},
       {"repo_type":"deb", "repo_url":"https://apt.repos.neuron.amazonaws.com/"}
     ],
-    "manifest_date": "10/25/2024",
-    "manifest_version": "2.20.1",
+    "manifest_date": "11/20/2024",
+    "manifest_version": "2.20.2",
     "latest_release": [
-      {"instance":"inf1", "version":"2.20.1"},
-      {"instance":"trn1", "version":"2.20.1"},
-      {"instance":"inf2", "version":"2.20.1"},
-      {"instance":"trn1n", "version":"2.20.1"}
+      {"instance":"inf1", "version":"2.20.2"},
+      {"instance":"trn1", "version":"2.20.2"},
+      {"instance":"inf2", "version":"2.20.2"},
+      {"instance":"trn1n", "version":"2.20.2"}
     ],
     "os_properties": [
       {"os":"ubuntu18", "default_python_version":"3.7"},
@@ -90,6 +90,56 @@
       {"name":"jax_neuronx","component":"Jax","category":"jax","package_type":"pip","use_cases":["inference"],"pin_major":"true"}
     ],
     "neuron_releases": [
+      {"neuron_version":"2.20.2", "packages": [
+        {"name":"aws-neuronx-collectives","version":"2.22.33.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-collectives","version":"2.12.35.0","supported_instances":["inf1"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-dkms","version":"2.18.20.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-gpsimd-customop-lib","version":"0.12.2.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-gpsimd-tools","version":"0.12.1.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-k8-plugin","version":"2.22.20.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-k8-scheduler","version":"2.22.20.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-oci-hook","version":"2.5.8.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-runtime-discovery","version":"2.9","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"aws-neuronx-runtime-lib","version":"2.12.23.0","supported_instances":["inf1"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-runtime-lib","version":"2.22.19.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
+        {"name":"aws-neuronx-tools","version":"2.19.0.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"dmlc_nnvm","version":"1.19.6.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"dmlc_topi","version":"1.19.6.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"dmlc_tvm","version":"1.19.6.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"inferentia_hwm","version":"1.17.6.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"jax_neuronx","version":"0.1.1","supported_instances":["trn1","inf2"],"supported_python_versions":["3.9"]},
+        {"name":"libneuronxla","version":"2.0.5347.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"libneuronxla","version":"0.5.3278","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"mx_neuron","version":"1.8.0.2.4.147.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"mxnet_neuron","version":"1.5.1.1.10.0.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"neuron-cc","version":"1.24.0.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"neuronperf","version":"1.8.93.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"neuronx-cc","version":"2.15.143.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"neuronx-cc-stubs","version":"2.15.143.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"neuronx_distributed","version":"0.9.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"neuronx_distributed_training","version":"1.0.1","supported_instances":["trn1"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"tensorboard-plugin-neuronx","version":"2.6.63.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
+        {"name":"tensorflow-model-server-neuronx","version":"2.10.1.2.12.2.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"tensorflow-model-server-neuronx","version":"2.9.3.2.12.2.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"tensorflow-model-server-neuronx","version":"2.8.4.2.12.2.0","supported_instances":["inf1","trn1","inf2"],"supported_python_versions":[]},
+        {"name":"tensorflow-neuron","version":"2.10.1.2.12.2.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"tensorflow-neuron","version":"2.8.4.2.12.2.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"tensorflow-neuron","version":"2.9.3.2.12.2.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},      
+        {"name":"tensorflow-neuronx","version":"2.10.1.2.1.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"tensorflow-neuronx","version":"2.8.4.2.1.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"tensorflow-neuronx","version":"2.9.3.2.1.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuron","version":"1.10.2.2.11.13.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuron","version":"1.11.0.2.11.13.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuron","version":"1.12.1.2.11.13.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuron","version":"1.13.1.2.11.13.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuron","version":"1.9.1.2.11.13.0","supported_instances":["inf1"],"supported_python_versions":["3.8","3.9","3.10"]},
+        {"name":"torch-neuronx","version":"1.13.1.1.16.0","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"torch-neuronx","version":"2.1.2.2.3.2","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"torch_xla","version":"1.13.1+torchneurong","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"torch_xla","version":"2.1.5","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"transformers-neuronx","version":"0.12.313","supported_instances":["trn1","inf2"],"supported_python_versions":["3.8","3.9","3.10","3.11"]},
+        {"name":"efa-installer","version":"na","supported_instances":["trn1"],"supported_python_versions":[]}
+      ]},
       {"neuron_version":"2.20.1", "packages": [
         {"name":"aws-neuronx-collectives","version":"2.22.26.0","supported_instances":["trn1","inf2"],"supported_python_versions":[]},
         {"name":"aws-neuronx-collectives","version":"2.12.35.0","supported_instances":["inf1"],"supported_python_versions":[]},
diff --git a/static/robots.txt b/static/robots.txt
@@ -1,5 +1,7 @@
 User-agent: *
 
+Disallow: /en/v2.20.2/
+
 Disallow: /en/v2.20.1/
 
 Disallow: /en/v2.20.0/
