Open
Conversation
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
nargokul
temporarily deployed
to
auto-approve
GitHub Actions
Inactive
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Bug 1: inference_config ignored for model customization deployments deploy() accepted inference_config with ResourceRequirements (accelerator count, memory, CPUs) but never passed it to _deploy_model_customization(). Fixed by forwarding inference_config through deploy() and extracting its values into the CreateInferenceComponent API call.
Bug 2: Incorrect compute requirements causing deployment failures _fetch_and_cache_recipe_config() used metadata memory values that exceeded SageMaker limits, passed accelerator counts to CPU instances, and used a wrong artifact path for fine-tuned models. Fixed by using safe defaults (1024 MB memory), dynamically querying EC2 for GPU detection/counts, stripping accelerators for CPU instances, and returning None artifact URL for fine-tuned models.
Bug 3: Base model evaluation uses fine-tuned model weights (sagemaker-train) The EvaluateBaseInferenceModel step in the LLM-as-Judge pipeline template included ModelPackageConfig with SourceModelPackageArn, causing it to load fine-tuned weights instead of the base model. Fixed by removing ModelPackageConfig from the base model evaluation step so it uses only the base model from the public hub.
More info: https://tiny.amazon.com/1icjlpkmh
The train integ tests seem to be failing with a ResourceLimitExceeded Error . We will need to track this fix separately .