fix(stablediffusion-ggml): Build with DSD CUDA, HIP and Metal flags #5236
+6
−6
Conversation
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
✅ Deploy Preview for localai ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
richiejp
force-pushed
the
cuda-fix
branch
2 times, most recently
from
d844e20 to
b20a86b
Compare
Signed-off-by: Richard Palethorpe <[email protected]>
mudler
approved these changes
Apr 24, 2025
|
Thank you again @richiejp ! |
Bloodis94
pushed a commit
to Bloodis94/LocalAI
that referenced
this pull request
Apr 24, 2025
…udler#5236) Signed-off-by: Richard Palethorpe <[email protected]>
mudler
added a commit
that referenced
this pull request
Apr 26, 2025
…n fedora without pulling X11 (#5246) * Update installation script for improved compatibility and clarity - Renamed VERSION to LOCALAI_VERSION to avoid conflicts with system variables. - Enhanced NVIDIA and CUDA repository installation for DNF5 compatibility. - Adjusted default Fedora version handling for CUDA installation. - Updated Docker image tag handling to use LOCALAI_VERSION consistently. - Improved logging messages for repository and LocalAI binary downloads. - Added a temporary bypass for nvidia-smi installation on Fedora Cloud Edition. * Enhance log functions with ANSI color formatting - Added ANSI escape codes for improved log styling: light blue for info, orange for warnings, and red for errors. - Updated all log functions (`info`, `warn`, `fatal`) to include bold and colored output. Signed-off-by: Alessandro Pirastru <[email protected]> * feat: Enhance log functions with ANSI color formatting - Added ANSI escape codes for improved log styling: light blue for info, orange for warnings, and red for errors. - Updated all log functions (`info`, `warn`, `fatal`) to include bold and colored output. Signed-off-by: Alessandro Pirastru <[email protected]> * chore: ⬆️ Update ggml-org/llama.cpp to `ecda2ec4b347031a9b8a89ee2efc664ce63f599c` (#5238) :arrow_up: Update ggml-org/llama.cpp Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: mudler <[email protected]> * fix(stablediffusion-ggml): Build with DSD CUDA, HIP and Metal flags (#5236) Signed-off-by: Richard Palethorpe <[email protected]> * feat(install): enhance script with choice functions and logs - Added custom `choice_info`, `choice_warn`, and `choice_fatal` functions for interactive input logging. - Adjusted Docker volume creation message for better clarity. - Included NVIDIA driver check log for improved feedback to users. - Added consistent logging before starting LocalAI Docker containers across configurations. Signed-off-by: Alessandro Pirastru <[email protected]> * feat(install): add Fedora NVIDIA driver installation option - Introduced a new function to install NVIDIA kernel drivers on Fedora using akmod packages. - Added user prompt to choose between installing drivers automatically or exiting for manual setup. - Integrated the new function into the existing Fedora-specific CUDA toolkit installation workflow. Signed-off-by: Alessandro Pirastru <[email protected]> * fix(install): correct repository ID for DNF5 configuration - Update repository ID from 'nome-repo' to 'nvidia-cuda' for DNF5. - Ensures the correct repository name matches expected configuration. - Fix prevents potential misconfiguration during installation process. Signed-off-by: Alessandro Pirastru <[email protected]> * feat(install): enhance NVIDIA driver handling on Fedora - fixed `install_cuda_driver_yum` function call in `install_fedora_nvidia_kernel_drivers` - Added `cuda-toolkit` for Fedora installations, as recommended by RPM Fusion. - Adjusted driver repository commands for compatibility with DNF5. - Improved URL and version handling for package manager installations. Signed-off-by: Alessandro Pirastru <[email protected]> * Refactor NVIDIA driver installation process in install.sh - Removed redundant empty lines for cleaner formatting. - Standardized URL formatting by removing unnecessary quotes around URLs. - Reverted logic by removing Fedora-specific exclusions for cuda-toolkit and using `cuda-drivers` universally. - Refined repository addition for `dnf` by explicitly setting `id` and `name` parameters for clarity and accuracy. - Fixed minor formatting inconsistencies in parameter passing. Signed-off-by: Alessandro Pirastru <[email protected]> * feat: Update NVIDIA module installation warning in install script - Clarified that Akmod installation may inhibit the reboot command. - Added a cautionary note to the warning to inform users of potential risks. Signed-off-by: Alessandro Pirastru <[email protected]> * Update NVIDIA driver installation warning message - Clarify prerequisites by noting the need for rpmfusion free/nonfree repos. - Improve formatting of the warning box for better readability. - Inform users that the script will install missing repos if necessary. Signed-off-by: Alessandro Pirastru <[email protected]> --------- Signed-off-by: Alessandro Pirastru <[email protected]> Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Signed-off-by: Richard Palethorpe <[email protected]> Co-authored-by: LocalAI [bot] <[email protected]> Co-authored-by: mudler <[email protected]> Co-authored-by: Richard Palethorpe <[email protected]>
1 task
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
Signed-off-by: Richard Palethorpe [email protected]
Description
This PR maybe fixes CUDA being broken for stablediffusion-ggml. Also HIP and Metal because none of them used the SD variables which set teh GGML ones and activate some ifdefs in SD.
Notes for Reviewers
I haven't tested this yet
Signed commits