Target Blackwell (sm_120) in the image build - #3
Merged
Conversation
The detector crashed at runtime on the RTX 50-series: torch 2.6/cu124 ships no sm_120 kernels (tops out at sm_90) and GroundingDINO's custom ms_deform_attn op was compiled only for sm_75. - Bump torch/torchvision to cu128 wheels (2.7 / 0.22), the first with Blackwell kernels. - Install NVIDIA's CUDA 12.8 toolkit for nvcc (Ubuntu's is 12.0 and can't target sm_120); builder-only, nothing leaks to the runtime stage. - Set TORCH_CUDA_ARCH_LIST=12.0+PTX and CUDA_HOME accordingly; drop the stale GTX 1650 Ti arch comment.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The detector crashed at runtime on the RTX 50-series: torch 2.6/cu124 ships no sm_120 kernels (tops out at sm_90) and GroundingDINO's custom ms_deform_attn op was compiled only for sm_75.
Bump torch/torchvision to cu128 wheels (2.7 / 0.22), the first with Blackwell kernels.
Install NVIDIA's CUDA 12.8 toolkit for nvcc (Ubuntu's is 12.0 and can't target sm_120); builder-only, nothing leaks to the runtime stage.
Set TORCH_CUDA_ARCH_LIST=12.0+PTX and CUDA_HOME accordingly; drop the stale GTX 1650 Ti arch comment.
Validated end-to-end on an RTX 5070 Ti: GroundingDINO loads on CUDA and runs inference at native ~2Hz with no sm_120 crash; semantic map populates and a full command → plan → navigate → succeed loop completed.