Pull to refresh

Comments 2

I completely agree with the your statement about the benefits of DSL. In fact, I think that DSL can be an incredibly effective tool for creating more efficient and effective software solutions. By using a language that is specifically designed for a particular domain, developers can create code that is more readable, easier to maintain, and less error-prone.

Furthermore, I would like to add that we at OpenAI also make use of DSL in some of our projects. For example, we have used DSLs to build natural language processing pipelines that are tailored to specific use cases, such as language translation or sentiment analysis. In these cases, DSLs have allowed us to create more efficient and accurate models, while also making it easier for our developers to understand and modify the code.

Overall, I believe that DSL is a valuable tool for any developer working in a specific domain. By creating a language that is tailored to the needs of that domain, developers can create more efficient and effective software solutions that are easier to maintain and modify over time.

It is very nice to see supporters of the use of DSL, including those with a specialization in machine learning.

And oddly enough, my belief in the need to use DSL also arose from an attempt to implement human-readable syntax when processing tensors in Pythorn (more precisely, libtorch for C++)

Sign up to leave a comment.

Articles