·
5 commits
to main
since this release
Version 3.0.1
New Features
llama4 scout model available for inference on Sagemaker
- The Meta Llama 4 Scout 17B 16E Instruct model is now available in N. Virginia (us-east-1) region for inferencing
- The model is currently launched on p4d.24xlarge. Suggested instances are ml.p5.48xlarge, ml.g6e.48xlarge, ml.p5.48xlarge, ml.p5en.48xlarge. More about it llama4 on JumpStart
Enhancements
Expert Evaluation
- Enabled Expert Evaluation button post completion of at least two experiments
- Multiple UI and label enhancements
Improved prompt file template for more accurate answer generation
Bug Fixes
- Downloaded experiment results csv file includes only relevant columns
- Sorting issue for projects has been fixed
- NaN issue for directional pricing has been fixed