fix: nss image, remove ffmpeg, force wandb version - #463
Conversation
612b189 to
87af3ec
Compare
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (5)
✅ Files skipped from review due to trivial changes (1)
🚧 Files skipped from review as they are similar to previous changes (3)
📝 WalkthroughWalkthroughPins ChangesDocker dependency and cleanup alignment
Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
|
8654d42 to
efe42ec
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
docker/Dockerfile.nmp-unsloth-training (1)
126-127: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueStale toolkit reference in commented block. Base is now 26.05 / CUDA 13.2, but this TODO still says NGC 26.02 and
/usr/local/cuda-13.1. Update so the future flash-attn re-enable doesn't target the wrong toolkit.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docker/Dockerfile.nmp-unsloth-training` around lines 126 - 127, The commented Flash Attention 2 TODO is stale and references the old NGC 26.02 torch and /usr/local/cuda-13.1, which no longer matches the current base image. Update the comment in the Dockerfile.nmp-unsloth-training block to reflect the current NGC 26.05 / CUDA 13.2 toolchain so the future flash-attn rebuild points to the correct toolkit. Use the existing Step 1d Flash Attention 2 comment as the anchor when editing.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pyproject.toml`:
- Line 299: The pydantic dependency entry is still unbounded on the upper end,
which lets lock refreshes pull in newer releases that break stainless
generation. Update the dependency declaration in the pyproject.toml package list
to add an upper bound or exact pin for pydantic, keeping the existing minimum
while preventing newer incompatible versions from being selected.
---
Nitpick comments:
In `@docker/Dockerfile.nmp-unsloth-training`:
- Around line 126-127: The commented Flash Attention 2 TODO is stale and
references the old NGC 26.02 torch and /usr/local/cuda-13.1, which no longer
matches the current base image. Update the comment in the
Dockerfile.nmp-unsloth-training block to reflect the current NGC 26.05 / CUDA
13.2 toolchain so the future flash-attn rebuild points to the correct toolkit.
Use the existing Step 1d Flash Attention 2 comment as the anchor when editing.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 73397f58-06ba-4f1e-bb28-9d93f0f275bb
⛔ Files ignored due to path filters (1)
uv.lockis excluded by!**/*.lock
📒 Files selected for processing (8)
docker-bake.hcldocker/Dockerfile.nmp-automodel-basedocker/Dockerfile.nmp-unsloth-trainingdocker/Dockerfile.safe-synthesizer-tasksdocker/automodel/no_override_requirements.txtdocker/scripts/cve-cleanup.shdocker/unsloth/no_override_requirements.txtpyproject.toml
✅ Files skipped from review due to trivial changes (1)
- docker/unsloth/no_override_requirements.txt
🚧 Files skipped from review as they are similar to previous changes (2)
- docker/Dockerfile.safe-synthesizer-tasks
- docker/scripts/cve-cleanup.sh
Signed-off-by: Matt Kornfield <mkornfield@nvidia.com>
Signed-off-by: Matt Kornfield <mkornfield@nvidia.com>
efe42ec to
b1c365d
Compare
Summary by CodeRabbit
Bug Fixes
Chores