Merge branch 'main' into base-cu1302-docs-update

Author: bhanutejagk

.github/workflows/pr-sglang.yml

@@ -19,7 +19,7 @@ env:
   SGLANG_VERSION: "0.5.8"
   PYTHON_VERSION: "py312"
   CUDA_VERSION: "cu129"
-  OS_VERSION: "ubuntu22.04"
+  OS_VERSION: "ubuntu24.04"
 
   # SGLang EC2
   PROD_EC2_IMAGE: sglang:0.5-gpu-py312-ec2

.gitignore

@@ -5,6 +5,7 @@ __pycache__
 .venv
 .ruff_cache
 .pytest_cache
+.vscode
 
 # Generated docs
 docs/index.md

.kiro/steering/docs.md

@@ -76,9 +76,9 @@ docs/src/
 
 - `constants.py` - Path constants, global variables, `GLOBAL_CONFIG`, `SITE_URL`, `README_PATH`, and `RELEASE_NOTES_REQUIRED_FIELDS`
 - `sorter.py` - Sorting tiebreaker functions: `platform_sorter`, `accelerator_sorter`, `repository_sorter`
-- `utils.py` - Utility functions: `load_yaml()`, `load_table_config()`, `load_jinja2()`, `render_table()`, `write_output()`, `parse_version()`, `clone_git_repository()`, `build_ecr_uri()`, `build_public_ecr_uri()`, `get_framework_order()`
+- `utils.py` - Utility functions: `load_yaml()`, `load_table_config()`, `load_jinja2()`, `render_table()`, `write_output()`, `parse_version()`, `clone_git_repository()`, `build_ecr_uri()`, `build_public_ecr_uri()`, `get_framework_order()`, `flatten_group_repos()`
 - `image_config.py` - `ImageConfig` class, image loaders (`load_repository_images`, `load_legacy_images`, `load_images_by_framework_group`), `sort_by_version`, `get_latest_image_uri`, `build_image_row`, `check_public_registry`
-- `generate.py` - `generate_index()`, `generate_support_policy()`, `generate_available_images()`, `generate_release_notes()`, `generate_all()`
+- `generate.py` - `generate_index()`, `generate_support_policy()`, `generate_available_images()`, `generate_release_notes()`, `generate_all()`, helpers: `_consolidate_framework_version()`, `_collapse_minor_versions()`
 - `macros.py` - MkDocs macros plugin integration
 - `hooks.py` - MkDocs hooks entry point
 
@@ -356,16 +356,19 @@ display_names:
   known_issues: "Known Issues"
 
 # Framework groups for support policy consolidation (lowercase keys)
+# Flat list: repos split directly to per-repo rows on mismatch
+# Nested dict: repos consolidate by sub-group first on mismatch
 framework_groups:
   pytorch:
-    - pytorch-training
-    - pytorch-inference
-    - pytorch-training-arm64
-    - pytorch-inference-arm64
-  tensorflow:
-    - tensorflow-training
-    - tensorflow-inference
-    - tensorflow-inference-arm64
+    pytorch-training:
+      - pytorch-training
+      - pytorch-training-arm64
+    pytorch-inference:
+      - pytorch-inference
+      - pytorch-inference-arm64
+  vllm:
+    - vllm
+    - vllm-arm64
 
 # Table order (controls order in available_images.md and support_policy.md)
 table_order:
@@ -376,20 +379,25 @@ table_order:
 
 ### Support Policy Consolidation
 
-The `framework_groups` configuration consolidates support policy rows by framework. Repositories in the same group are combined into a single row using the framework name (e.g., "PyTorch").
+The `framework_groups` configuration consolidates support policy rows by framework. The generation follows a 3-step flow:
 
-**Version Display:**
+**Validation (at load time):** `load_repository_images` validates that images sharing the same full version within a single repository have identical GA/EOP dates. Raises `ValueError` if not (this is a data bug).
 
-- Images with the same major.minor version (e.g., `2.6.0` and `2.6.1`) are consolidated into a single row displayed as `2.6` if they have identical GA/EOP dates
-- If patch versions have different GA/EOP dates, each is displayed separately with full version (e.g., `2.6.0`, `2.6.1`) and a warning is logged
+**Step 1 — Group by full version:** All images in a framework group are grouped by full version (e.g., `2.6.0`), deduplicated per repository (one representative image per repo since intra-repo consistency is guaranteed).
 
-**Requirements:**
+**Step 2 — Hierarchical consolidation** via `_consolidate_framework_version()`: For each full version, tries three levels of date agreement, stopping at the first that succeeds:
+
+1. Framework group — all repos agree → single row (e.g., "PyTorch")
+1. Sub-group — repos within a sub-group agree → one row per sub-group (e.g., "PyTorch Inference"). Only applies to nested dict groups in `framework_groups`.
+1. Per-repo — no agreement → one row per repository using its individual display name
+
+**Step 3 — Major.minor collapse:** Non-split entries are grouped by major.minor. If all full versions within a major.minor share the same dates, they collapse into a single row displayed as the major.minor (e.g., `2.6`). Collapse is skipped for any major.minor that has split (per-repo) rows.
+
+**Behavior:**
 
-- All repositories in a group that have a given full version (X.Y.Z) must have identical GA/EOP dates
 - Missing versions in some repositories are allowed (only present repos are consolidated)
-- A `ValueError` is raised if dates differ within a group for the same full version
 
-To add a new framework group, add an entry to `framework_groups` with the framework name as key and list of repositories as value.
+To add a new framework group, add an entry to `framework_groups` with the framework name as key and list of repositories as value. Use nested dict format if you need intermediate sub-group consolidation when dates differ.
 
 ### Reordering Tables and Columns
 

README.md

@@ -19,6 +19,12 @@ ______________________________________________________________________
 - **[2025/11/17]** Released first [SGLang DLCs](https://gallery.ecr.aws/deep-learning-containers/sglang)
   - SageMaker: `public.ecr.aws/deep-learning-containers/sglang:0.5.5-gpu-py312`
 
+### 📢 Support Updates
+
+- **[2026/02/10]** Extended support for PyTorch 2.6 Inference containers until June 30, 2026
+  - PyTorch 2.6 Inference images will continue to receive security patches and updates through end of June 2026
+  - For complete framework support timelines, see our [Support Policy](https://aws.github.io/deep-learning-containers/reference/support_policy/)
+
 ### 🎉 Hot Off the Press
 
 - 🌐

docs/DEVELOPMENT.md

@@ -65,7 +65,7 @@ See `docs/src/data/template/image-template.yml` for all available fields.
 ### Step 2: Regenerate
 
 ```bash
-cd docs/src && python main.py --verbose
+python docs/src/main.py --verbose
 ```
 
 ---
@@ -76,15 +76,19 @@ Add `ga` and `eop` fields to image configs for repositories that appear in suppo
 
 ```yaml
 ga: "2025-10-15"    # General Availability date
-eop: "2026-10-15"   # End of Patch date
+eop: "2035-10-15"   # End of Patch date
 ```
 
 **Version Consolidation:**
 
-- Images with the same major.minor version (e.g., `2.6.0` and `2.6.1`) are consolidated into a single row displayed as `2.6` if they have identical GA/EOP dates
-- If patch versions have different GA/EOP dates, each is displayed separately with full version (e.g., `2.6.0`, `2.6.1`) and a warning is logged
+- Images with the same major.minor version and identical GA/EOP dates are consolidated into a single row displayed as `2.6` with the framework group name (e.g., "PyTorch")
+- If the same version has different GA/EOP dates across repository types (e.g., training vs inference), separate rows are created showing the specific repository type: "PyTorch Training" and "PyTorch Inference"
+- ARM64 variants are automatically consolidated with their base repository
+- If patch versions within the same repository have different GA/EOP dates, each is displayed separately with full version (e.g., `2.6.0`, `2.6.1`) and a warning is logged
 
-**Validation:** All images in the same framework group with the same full version (X.Y.Z) must have identical GA/EOP dates.
+**Flexibility:** Repositories in the same framework group (e.g., pytorch-training and pytorch-inference) can have different GA/EOP dates for the same version. The system will automatically create separate rows showing the specific repository type when dates differ.
+
+**Example:** If PyTorch 2.6 Training has EOP 2025-10-15 but PyTorch 2.6 Inference has EOP 2026-10-15, the support policy table will show two separate rows with "PyTorch Training" and "PyTorch Inference" in the Framework column.
 
 ---
 
@@ -229,4 +233,4 @@ For any changes required to the tutorial pages, create a new PR in
 | "Display name not found" | Add repository to `display_names` in `global.yml` |
 | "Inconsistent dates" | Ensure all images in same framework group/version have identical GA/EOP |
 | Images not appearing | Check repository is in `table_order` |
-| Release notes not generating | Ensure `announcements` and `packages` fields are present |
+| Release notes not generating | Ensure `announcements` and `packages` fields are present |

docs/README.md

@@ -18,23 +18,23 @@ uv pip install -r docs/requirements.txt
 
 Run the generation system without serving:
 ```bash
-cd docs/src && python main.py --verbose
+python docs/src/main.py --verbose
 ```
 This clones the `tutorials/` repository and generates markdown files in `reference/` and `releasenotes/` directories.
 
 Generation flags:
 ```bash
-python main.py --available-images-only
-python main.py --support-policy-only
-python main.py --release-notes-only
-python main.py --index-only
+python docs/src/main.py --available-images-only
+python docs/src/main.py --support-policy-only
+python docs/src/main.py --release-notes-only
+python docs/src/main.py --index-only
 ```
 
 ### Serving
 
 Use `mkdocs serve` to generate documentation and serve the website:
 ```bash
-cd docs && mkdocs serve
+mkdocs serve
 ```
 The site is typically available at `http://127.0.0.1:8000/deep-learning-containers/` - check the command output for the actual URL.
 

docs/src/data/pytorch-inference-arm64/2.6-cpu-ec2.yml

@@ -5,7 +5,7 @@ python: py312
 os: ubuntu22.04
 platform: ec2
 ga: "2025-01-29"
-eop: "2026-01-29"
+eop: "2026-06-30"
 public_registry: true
 
 tags:

docs/src/data/pytorch-inference-arm64/2.6-cpu-sagemaker.yml

@@ -5,7 +5,7 @@ python: py312
 os: ubuntu22.04
 platform: sagemaker
 ga: "2025-01-29"
-eop: "2026-01-29"
+eop: "2026-06-30"
 public_registry: true
 
 tags:

docs/src/data/pytorch-inference-arm64/2.6-gpu-ec2.yml

@@ -6,7 +6,7 @@ cuda: cu124
 os: ubuntu22.04
 platform: ec2
 ga: "2025-01-29"
-eop: "2026-01-29"
+eop: "2026-06-30"
 public_registry: true
 
 tags:

docs/src/data/pytorch-inference/2.6-cpu-ec2.yml

@@ -5,7 +5,7 @@ python: py312
 os: ubuntu22.04
 platform: ec2
 ga: "2025-01-29"
-eop: "2026-01-29"
+eop: "2026-06-30"
 public_registry: true
 
 tags: