What is Salesforce's CodeRL?
CodeRL is a general framework that integrates pretrained LMs and RL holistically for program synthesis
I’m really trying my best to follow the intersection of A.I. and Code. As some of you know I’m really bullish on Salesforce Research, Einstein and the quality of papers they are now involved with.
While on AI Supremacy I’ve covered a lot of DeepMind, OpenAI and Meta AI, it’s time to also cover more from other sources including Salesforce, China and MIT.
Excited to introduce CodeRL: a novel code generation framework as a whole new way of building SOTA AI Coding systems by combining Pretrained Models and Deep Reinforcement Learning (RL). Code & models are open-source: Code: github.com/salesforce/Cod… Paper: arxiv.org/abs/2207.01780
You can find them on GitHub or Hugging Face.
Our CodeGen models are now available at @huggingface! (Model size variants: 350M, 2B, 6B, and 16B.) Clone the latest transformers repository and try it out! Paper: arxiv.org/abs/2203.13474 Models: huggingface.co/models?search=…
GET A BACK STAGE PASS ACCESS TO PAID CONTENT
Keep reading with a 7-day free trial
Subscribe to AI Supremacy to keep reading this post and get 7 days of free access to the full post archives.