SPARKNLP 1034 implement starcoder2 for causal lm #14358
Merged
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Outdated suggestions cannot be applied.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.
This PR introduces StarCoder2
Description
This pull request adds support for the StarCoder2 model in the Spark NLP library. StarCoder2 is a Transformer model designed specifically for code generation and understanding. It features 3 billion parameters and is trained on a diverse dataset including multiple programming languages, making it highly versatile for various coding tasks. The new model provides enhanced functionality for code completion, generation, and understanding.
Motivation and Context
The inclusion of StarCoder2 addresses the need for a robust code-focused language model within Spark NLP. This model will significantly improve the library's capabilities in code generation and understanding tasks, offering developers a powerful tool for software development and data science projects.
How Has This Been Tested?
The StarCoder2 model has been tested using various unit and integration tests to ensure its proper functionality within the Spark NLP framework. Tests were conducted to verify the model's ability to perform code completion and generation tasks accurately. Additionally, performance benchmarks were compared to existing models to ensure its efficacy.
Screenshots (if appropriate):
Types of changes
Checklist: