Bug fixes 01302026#23
Merged
Merged
Conversation
Changes: 1. Added OpenMPI/OpenBLAS libraries in builder stage - Installs libopenmpi3, libopenmpi-dev, libopenblas0, libopenblas-dev - Required dependencies for PyTorch on Jetson 2. Fixed PyTorch installation for Jetson - Before: Used broken pypi.jetson-ai-lab.dev/jp6/cu126 index (wrong CUDA version) - After: Downloads PyTorch 2.3.0 wheel directly from NVIDIA Box - Pins torchvision==0.18.0 for compatibility 3. Added build-time verification - Prints PyTorch version and CUDA build version - Informational only (no assertion since GPU not available during build) 4. Added runtime libraries - Installs libopenmpi3, libopenblas0 in runtime stage - Required for PyTorch to load at runtime
- Auto-detects ROS2 distribution (Humble/Jazzy/Iron) - Installs all required ROS2 packages (cv-bridge, rviz2, image-publisher, etc.) - Installs Python dependencies (PyTorch, transformers, OpenCV, etc.) - Installs Depth Anything 3 from ByteDance - Builds ROS2 workspace automatically - Downloads sample images - Handles Jetson ARM64 architecture
- Auto-sources ROS2 and builds workspace if needed - Auto-installs missing ROS2 packages (rviz2, image-publisher, etc.) - Launches 5 gnome-terminal windows for comprehensive demo: 1. Depth estimation node with image publisher 2. RViz2 visualization 3. Topic monitoring with interactive menu 4. Parameter inspection 5. Additional topics (confidence, colored depth) - Correct namespace handling (/test/depth_anything_3) - Comprehensive logging to /tmp/da3_demo_logs/ - Python dependency checks before launch - Clean shutdown with Ctrl+C
- Add Quick Install section with one-command setup - Document install_dependencies.sh script - Update Table of Contents with new sections - Add troubleshooting info for demo script - Document logging locations
- Indoor, outdoor, and objects test images - Used by demo script for depth estimation
…p checks and instructions - Handle dustynv Jetson image ROS paths in entrypoint
…vert CRLF to LF for Python and shell scripts
…registry instead of direct DepthAnything3Net - Map HuggingFace model names to registry names
… - Use correct ROS setup path for dustynv images
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces major improvements for NVIDIA Jetson support, validates TensorRT 10.3 performance, and adds robust hardware detection, model setup, and testing infrastructure. It also resolves critical compatibility issues with OpenCV, PyTorch, and ROS2 on Jetson platforms, and updates the Docker build process for seamless deployment across CPU, CUDA, and Jetson ARM64 environments.
Jetson and TensorRT Support
Dockerfileusingdustynv/ros:humble-pytorch-l4t-${L4T_VERSION}, with full OpenCV 4.8.1 CUDA support and source builds forcv_bridgeandimage_geometryto resolve ROS/OpenCV conflicts.Hardware Detection and Model Setup
jetson_detector.pyfor automatic Jetson module, GPU, VRAM, JetPack, and CUDA detection.Testing and CI Improvements
.gitignoreentries for new files and enforced LF line endings for shell scripts to avoid Windows CRLF issues.Documentation and Developer Guidance
CLAUDE.mdwith detailed build, test, and deployment instructions, architecture overview, coding standards, and specialized agent usage for Jetson and NVIDIA troubleshooting.CHANGELOG.mdwith comprehensive release notes, critical findings, and resolved issues for Jetson and TensorRT support.Dependency and Build Fixes
--no-depsand manually managed ARM64 dependencies, patching for missing modules at runtime. [1] [2]References:
[1] [2] [3] [4] [5]## Description